From b9c41dc60a38371b8248c21be08e91ed65ebcf3d Mon Sep 17 00:00:00 2001 From: Dominik Date: Thu, 21 May 2026 16:03:54 -0400 Subject: [PATCH 1/2] feat: add GTM AI marketplace skills --- README.md | 20 +++++++++++ skills/account-prioritization/SKILL.md | 42 +++++++++++++++++++++++ skills/build-tam/SKILL.md | 42 +++++++++++++++++++++++ skills/business-case-builder/SKILL.md | 42 +++++++++++++++++++++++ skills/campaign-builder/SKILL.md | 42 +++++++++++++++++++++++ skills/champion-tracking/SKILL.md | 42 +++++++++++++++++++++++ skills/clay-migration/SKILL.md | 42 +++++++++++++++++++++++ skills/competitive-battlecard/SKILL.md | 42 +++++++++++++++++++++++ skills/crm-cleanup/SKILL.md | 42 +++++++++++++++++++++++ skills/deal-qualification/SKILL.md | 42 +++++++++++++++++++++++ skills/deal-risk-scorer/SKILL.md | 42 +++++++++++++++++++++++ skills/demo-scripting/SKILL.md | 42 +++++++++++++++++++++++ skills/discovery-brief/SKILL.md | 42 +++++++++++++++++++++++ skills/icp-definition/SKILL.md | 42 +++++++++++++++++++++++ skills/pipeline-hygiene/SKILL.md | 42 +++++++++++++++++++++++ skills/procurement-prep/SKILL.md | 42 +++++++++++++++++++++++ skills/renewal-risk-brief/SKILL.md | 42 +++++++++++++++++++++++ skills/stakeholder-mapper/SKILL.md | 42 +++++++++++++++++++++++ skills/territory-planning/SKILL.md | 42 +++++++++++++++++++++++ skills/trigger-based-list-build/SKILL.md | 42 +++++++++++++++++++++++ skills/waterfall-enrich/SKILL.md | 43 ++++++++++++++++++++++++ 21 files changed, 861 insertions(+) create mode 100644 skills/account-prioritization/SKILL.md create mode 100644 skills/build-tam/SKILL.md create mode 100644 skills/business-case-builder/SKILL.md create mode 100644 skills/campaign-builder/SKILL.md create mode 100644 skills/champion-tracking/SKILL.md create mode 100644 skills/clay-migration/SKILL.md create mode 100644 skills/competitive-battlecard/SKILL.md create mode 100644 skills/crm-cleanup/SKILL.md create mode 100644 skills/deal-qualification/SKILL.md create mode 100644 skills/deal-risk-scorer/SKILL.md create mode 100644 skills/demo-scripting/SKILL.md create mode 100644 skills/discovery-brief/SKILL.md create mode 100644 skills/icp-definition/SKILL.md create mode 100644 skills/pipeline-hygiene/SKILL.md create mode 100644 skills/procurement-prep/SKILL.md create mode 100644 skills/renewal-risk-brief/SKILL.md create mode 100644 skills/stakeholder-mapper/SKILL.md create mode 100644 skills/territory-planning/SKILL.md create mode 100644 skills/trigger-based-list-build/SKILL.md create mode 100644 skills/waterfall-enrich/SKILL.md diff --git a/README.md b/README.md index fd77f7e..68b3d6a 100644 --- a/README.md +++ b/README.md @@ -64,20 +64,40 @@ git clone https://github.com/Zoominfo/zoominfo-mcp-plugin.git | Skill | Description | |---|---| +| `account-prioritization` | Rank accounts by ICP fit, intent, trigger signals, whitespace, and explainable next-best action. | | `account-research` | Produce an account intelligence brief with firmographics, relationship context, intent, news, and next actions | | `build-list` | Build targeted account or contact lists from natural-language criteria | +| `build-tam` | Build a market or territory TAM using ZoomInfo company search counts, reusable ICP filters, sample accounts, and assumption notes. | +| `business-case-builder` | Draft a buyer-specific business case using account scale, pain hypotheses, relevant signals, stakeholders, and measurable outcome assumptions. | | `buying-committee` | Map decision-makers, influencers, champions, and coverage gaps at a target account | +| `campaign-builder` | Turn an ICP, trigger, or product motion into a segmented campaign audience with personas, proof points, messaging angles, and activation fields. | +| `champion-tracking` | Track champion movement, title changes, departures, influence, and replacement contacts across target or customer accounts. | +| `clay-migration` | Convert Clay-style enrichment tables and workflows into governed ZoomInfo MCP skill runs with equivalent outputs and fewer manual provider chains. | +| `competitive-battlecard` | Build an account- or competitor-specific battlecard from firmographics, news, scoops, intent, buying committee, and positioning evidence. | | `competitor-analysis` | Create fact-led competitor briefs using ZoomInfo data plus public context | +| `crm-cleanup` | Audit and repair CRM account/contact data using ZoomInfo matching, confidence scoring, duplicate detection, and safe writeback recommendations. | +| `deal-qualification` | Qualify an account or opportunity using firmographics, buyer fit, urgency signals, stakeholder coverage, and data confidence. | +| `deal-risk-scorer` | Score deal risk using account context, stakeholder coverage, trigger evidence, competitor signals, and missing buying-committee roles. | +| `demo-scripting` | Create account-specific demo talk tracks using company context, persona priorities, trigger signals, and competitive angles. | +| `discovery-brief` | Prepare a discovery-call brief with account context, buyer hypotheses, recent signals, personas, questions, and talk tracks. | | `enrich-company` | Look up company profiles, firmographics, financials, structure, and growth signals | | `enrich-contact` | Look up professional contact profiles, title, department, contact data, and accuracy signals | | `find-similar` | Find lookalike companies or contacts based on a reference account or person | +| `icp-definition` | Derive a reusable ICP from target accounts, winners, hypotheses, or market criteria using ZoomInfo firmographic and signal data. | | `meeting-prep` | Prepare for upcoming calls with account, attendee, relationship, and talking-point context | | `personalize-email` | Draft outreach grounded in account, contact, intent, and trigger signals | +| `pipeline-hygiene` | Inspect open pipeline for stale deals, bad data, missing stakeholders, weak timing signals, and account-risk indicators. | +| `procurement-prep` | Prepare for procurement by mapping stakeholders, account structure, risk signals, vendor/contract clues, and negotiation angles. | | `recommend-contacts` | Get AI-ranked contact recommendations at a target company | +| `renewal-risk-brief` | Assess renewal or expansion risk using account signals, champion status, stakeholder gaps, growth/contraction clues, and next actions. | | `score-accounts` | Prioritize accounts by ICP fit, intent, trigger signals, and explainable scoring | | `score-leads` | Rank inbound leads by fit, urgency, verified contact data, and recommended action | +| `stakeholder-mapper` | Map stakeholders at a target account by role, seniority, likely influence, buying-committee coverage, and recommended outreach order. | | `tam-sizer` | Size a market or territory and produce a reusable ICP filter set | | `tech-stack-snapshot` | Summarize detected technologies, displacement angles, and integration plays | +| `territory-planning` | Design or rebalance territories using ZoomInfo account counts, fit, coverage, trigger density, and rep-actionable segments. | +| `trigger-based-list-build` | Build account or contact lists around buying triggers such as intent, scoops, news, funding, hiring, technology, and leadership changes. | +| `waterfall-enrich` | Enrich company and contact records with ZoomInfo-first matching, field-level confidence, provenance, unresolved rows, and activation-ready output. | ## Project Structure diff --git a/skills/account-prioritization/SKILL.md b/skills/account-prioritization/SKILL.md new file mode 100644 index 0000000..878619d --- /dev/null +++ b/skills/account-prioritization/SKILL.md @@ -0,0 +1,42 @@ +--- +name: account-prioritization +description: Rank accounts by ICP fit, intent, trigger signals, whitespace, and explainable next-best action. +--- + +# Account Prioritization + +Rank accounts by ICP fit, intent, trigger signals, whitespace, and explainable next-best action. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Capture the prioritization goal**: new logo, expansion, renewal, competitive takeout, territory focus, or campaign audience. +2. **Normalize the input account list or build one with `search_companies`; resolve all filters with `lookup`.** +3. **Enrich accounts with firmographics, technologies, hierarchy, growth, and department data via `enrich_companies`.** +4. **Pull intent, scoops, and news where relevant; keep only signals that map to the prioritization goal.** +5. **Score accounts with transparent factors**: ICP fit, timing/urgency, reachable personas, data confidence, and strategic value. +6. **Assign each account to an action tier with a concise why-now rationale and recommended owner action.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Ranked account list with score breakdown, evidence, confidence, action tier, and next-best action. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/build-tam/SKILL.md b/skills/build-tam/SKILL.md new file mode 100644 index 0000000..0937a3a --- /dev/null +++ b/skills/build-tam/SKILL.md @@ -0,0 +1,42 @@ +--- +name: build-tam +description: Build a market or territory TAM using ZoomInfo company search counts, reusable ICP filters, sample accounts, and assumption notes. +--- + +# Build TAM + +Build a market or territory TAM using ZoomInfo company search counts, reusable ICP filters, sample accounts, and assumption notes. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Translate the user request into an explicit ICP**: industries, geography, employee count, revenue, technologies, growth/funding, ownership, and exclusions. +2. **Use `lookup` for every taxonomy-backed field before search. For industries, fuzzy-match each term separately and prefer specific sub-industries over broad parents.** +3. **Run `search_companies` with `pageSize**: 1` to capture `meta.totalResults` for the headline TAM count. +4. **Run segmented counts for the dimensions that matter most**: region, employee band, industry, revenue band, or technology. +5. **Pull representative accounts with `search_companies` using sorted views such as `-employeeCount` or `-revenue`; label samples as illustrative, not a census.** +6. **Stress-test assumptions**: run a broader and narrower variant and show sensitivity of TAM to the highest-leverage filters. + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: TAM estimate, applied filters, sensitivity ranges, segment breakdowns, sample accounts, exclusions, and reusable ICP filter set. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/business-case-builder/SKILL.md b/skills/business-case-builder/SKILL.md new file mode 100644 index 0000000..5d352fd --- /dev/null +++ b/skills/business-case-builder/SKILL.md @@ -0,0 +1,42 @@ +--- +name: business-case-builder +description: Draft a buyer-specific business case using account scale, pain hypotheses, relevant signals, stakeholders, and measurable outcome assumptions. +--- + +# Business Case Builder + +Draft a buyer-specific business case using account scale, pain hypotheses, relevant signals, stakeholders, and measurable outcome assumptions. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify buyer persona, product, desired outcome, account, and any known metric assumptions.** +2. **Enrich company data to size the opportunity**: employee count, revenue, locations, departments, growth, technologies, and hierarchy. +3. **Pull news, scoops, and intent to support urgency and business relevance.** +4. **Map stakeholders who will care about financial, operational, technical, and risk outcomes.** +5. **Build the case with explicit assumptions; use ranges when inputs are uncertain and do not fabricate ROI numbers.** +6. **Recommend proof points or data the seller should collect next to tighten the case.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Business case outline with account context, value drivers, stakeholder lens, quantified assumptions, risk notes, and next data asks. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/campaign-builder/SKILL.md b/skills/campaign-builder/SKILL.md new file mode 100644 index 0000000..f77138a --- /dev/null +++ b/skills/campaign-builder/SKILL.md @@ -0,0 +1,42 @@ +--- +name: campaign-builder +description: Turn an ICP, trigger, or product motion into a segmented campaign audience with personas, proof points, messaging angles, and activation fields. +--- + +# Campaign Builder + +Turn an ICP, trigger, or product motion into a segmented campaign audience with personas, proof points, messaging angles, and activation fields. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify campaign goal, offer, target segment, exclusion rules, and success metric. If absent, assume outbound prospecting and state it.** +2. **Build the account universe with `search_companies`; resolve filters through `lookup` first.** +3. **Identify target personas using `search_contacts`, `get_recommended_contacts`, or buying-committee logic depending on whether the account list is known.** +4. **Enrich account context with `enrich_companies`, `enrich_intent`, `enrich_scoops`, and `enrich_news` for segmentation and proof points.** +5. **Create practical segments**: fit tier, trigger type, persona, industry, company size, and activation priority. +6. **Draft campaign angles that cite observed data only; never invent initiatives, tools, or pains not supported by evidence.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Segmented campaign audience, persona map, proof points, messaging angles, exclusion logic, and export schema. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/champion-tracking/SKILL.md b/skills/champion-tracking/SKILL.md new file mode 100644 index 0000000..db948bd --- /dev/null +++ b/skills/champion-tracking/SKILL.md @@ -0,0 +1,42 @@ +--- +name: champion-tracking +description: Track champion movement, title changes, departures, influence, and replacement contacts across target or customer accounts. +--- + +# Champion Tracking + +Track champion movement, title changes, departures, influence, and replacement contacts across target or customer accounts. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Identify known champions by email, name, company, CRM/contact ID, or supplied list. Preserve original account association.** +2. **Resolve each person with `enrich_contacts`; fall back to `search_contacts` when needed and label ambiguous matches.** +3. **Check current title, company, management level, department, contact accuracy, and employment status from ZoomInfo fields.** +4. **If a champion moved, enrich the new company and classify the opportunity**: referral, expansion, new-logo, or monitor. +5. **At the original account, use `search_contacts` or recommended contacts to identify replacement champions and coverage gaps.** +6. **Prioritize by seniority, prior relationship value, account fit, and actionability.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Champion movement tracker with status, new company, action recommendation, replacement contacts, and confidence. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/clay-migration/SKILL.md b/skills/clay-migration/SKILL.md new file mode 100644 index 0000000..95e5967 --- /dev/null +++ b/skills/clay-migration/SKILL.md @@ -0,0 +1,42 @@ +--- +name: clay-migration +description: Convert Clay-style enrichment tables and workflows into governed ZoomInfo MCP skill runs with equivalent outputs and fewer manual provider chains. +--- + +# Clay Migration + +Convert Clay-style enrichment tables and workflows into governed ZoomInfo MCP skill runs with equivalent outputs and fewer manual provider chains. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Ask for or infer the current Clay table purpose**: account list, contact list, enrichment, scoring, personalization, or campaign prep. +2. **Inventory current columns, formulas, provider calls, and downstream activation fields; separate must-have outputs from nice-to-have fields.** +3. **Map Clay provider steps to ZoomInfo MCP steps**: `search_companies`, `search_contacts`, `enrich_companies`, `enrich_contacts`, `enrich_news`, `enrich_scoops`, `enrich_intent`, and recommendation tools where relevant. +4. **Resolve all ZoomInfo filter fields with `lookup`; document any fields with no clean ZoomInfo equivalent.** +5. **Design the replacement run as a sequence**: input normalization, entity resolution, enrichment, scoring/routing, QA, export/writeback. +6. **Produce a migration plan with parity gaps, confidence thresholds, manual-review queues, and quick wins that can ship first.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Migration blueprint plus replacement ZoomInfo workflow, field mapping, parity gaps, QA checks, and first-run test plan. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/competitive-battlecard/SKILL.md b/skills/competitive-battlecard/SKILL.md new file mode 100644 index 0000000..1fc210e --- /dev/null +++ b/skills/competitive-battlecard/SKILL.md @@ -0,0 +1,42 @@ +--- +name: competitive-battlecard +description: Build an account- or competitor-specific battlecard from firmographics, news, scoops, intent, buying committee, and positioning evidence. +--- + +# Competitive Battlecard + +Build an account- or competitor-specific battlecard from firmographics, news, scoops, intent, buying committee, and positioning evidence. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify whether the battlecard is competitor-led, account-led, or deal-led, and what decision it should support.** +2. **Resolve competitor and/or account company IDs; enrich company profiles and relevant firmographics.** +3. **Pull `enrich_news`, `enrich_scoops`, and `enrich_intent` for relevant entities; triage only signals connected to the competitive motion.** +4. **Map likely buyer personas and stakeholders at the target account when deal-specific.** +5. **Synthesize competitive strengths, vulnerabilities, likely objections, displacement triggers, and proof points.** +6. **Separate verified facts from positioning hypotheses and call out stale or low-confidence evidence.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Battlecard with competitor context, account-specific angles, objections, traps to avoid, proof points, and confidence notes. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/crm-cleanup/SKILL.md b/skills/crm-cleanup/SKILL.md new file mode 100644 index 0000000..fcdee19 --- /dev/null +++ b/skills/crm-cleanup/SKILL.md @@ -0,0 +1,42 @@ +--- +name: crm-cleanup +description: Audit and repair CRM account/contact data using ZoomInfo matching, confidence scoring, duplicate detection, and safe writeback recommendations. +--- + +# CRM Cleanup + +Audit and repair CRM account/contact data using ZoomInfo matching, confidence scoring, duplicate detection, and safe writeback recommendations. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Classify the cleanup request**: stale fields, missing fields, duplicates, bad domains, inactive contacts, title drift, account hierarchy, or territory/segment hygiene. +2. **Preserve source CRM IDs and current values. Never overwrite blindly; compare current CRM data to ZoomInfo-enriched values.** +3. **Resolve companies and contacts using stable identifiers first**: domain, email, ZoomInfo ID, then name + company fallback. +4. **Use `enrich_companies` and `enrich_contacts` to fill target fields; use contact accuracy and match strength to assign confidence.** +5. **Detect duplicates by shared domain, company IDs, email, normalized names, parent/subsidiary relationships, and conflicting firmographics.** +6. **Output explicit writeback actions**: update, append, merge-review, suppress, or no-change. + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Cleanup report with proposed field updates, confidence, duplicate/merge queues, suppressions, and safe writeback file. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/deal-qualification/SKILL.md b/skills/deal-qualification/SKILL.md new file mode 100644 index 0000000..8f9a38b --- /dev/null +++ b/skills/deal-qualification/SKILL.md @@ -0,0 +1,42 @@ +--- +name: deal-qualification +description: Qualify an account or opportunity using firmographics, buyer fit, urgency signals, stakeholder coverage, and data confidence. +--- + +# Deal Qualification + +Qualify an account or opportunity using firmographics, buyer fit, urgency signals, stakeholder coverage, and data confidence. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify qualification framework if supplied; otherwise use fit, pain/trigger, stakeholder access, timing, and data confidence.** +2. **Resolve the account and enrich core firmographics, technologies, company structure, and growth indicators.** +3. **Pull intent, scoops, and news to assess urgency; do not substitute generic industry assumptions for account-specific evidence.** +4. **Map relevant stakeholders and contactability using `search_contacts` and recommended contacts.** +5. **Score each qualification dimension with evidence and uncertainty; identify which missing fact would change the decision.** +6. **Recommend qualify, nurture, disqualify, or investigate with one next action.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Qualification scorecard with evidence, gaps, decision recommendation, and next action. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/deal-risk-scorer/SKILL.md b/skills/deal-risk-scorer/SKILL.md new file mode 100644 index 0000000..01e06df --- /dev/null +++ b/skills/deal-risk-scorer/SKILL.md @@ -0,0 +1,42 @@ +--- +name: deal-risk-scorer +description: Score deal risk using account context, stakeholder coverage, trigger evidence, competitor signals, and missing buying-committee roles. +--- + +# Deal Risk Scorer + +Score deal risk using account context, stakeholder coverage, trigger evidence, competitor signals, and missing buying-committee roles. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Start from named opportunity/account context and clarify deal stage, close date, product, known stakeholders, and primary concern if available.** +2. **Resolve the account via ZoomInfo ID, domain, or `search_companies`; enrich with `enrich_companies`.** +3. **Pull relevant news, scoops, and intent; treat absence of signal as uncertainty, not proof of no risk.** +4. **Map stakeholders and gaps using `search_contacts` and recommended contacts for deal acceleration.** +5. **Assess risks across five dimensions**: account fit, timing, stakeholder coverage, competitive pressure, and data confidence. +6. **Produce a mitigation plan with one owner-ready action per material risk.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Deal risk scorecard with evidence, confidence, stakeholder gaps, top risks, and mitigation actions. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/demo-scripting/SKILL.md b/skills/demo-scripting/SKILL.md new file mode 100644 index 0000000..bb73340 --- /dev/null +++ b/skills/demo-scripting/SKILL.md @@ -0,0 +1,42 @@ +--- +name: demo-scripting +description: Create account-specific demo talk tracks using company context, persona priorities, trigger signals, and competitive angles. +--- + +# Demo Scripting + +Create account-specific demo talk tracks using company context, persona priorities, trigger signals, and competitive angles. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Capture product/demo objective, account, attendees/personas, and desired outcome.** +2. **Enrich the company and attendees; map each attendee to likely role, priorities, and objections.** +3. **Pull recent news, scoops, intent, technology, and growth signals that can anchor the demo narrative.** +4. **Translate evidence into demo moments**: opening hook, workflow path, proof points, persona-specific benefits, and objection handling. +5. **Avoid pretending to know internal pain. Use external signals as hypotheses and phrase them as discovery-confirmed paths.** +6. **End with a crisp close plan**: success criteria, stakeholder follow-up, and next artifact. + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Demo script with opening, persona-specific talk tracks, evidence-backed proof points, objections, and close plan. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/discovery-brief/SKILL.md b/skills/discovery-brief/SKILL.md new file mode 100644 index 0000000..afdf7f7 --- /dev/null +++ b/skills/discovery-brief/SKILL.md @@ -0,0 +1,42 @@ +--- +name: discovery-brief +description: Prepare a discovery-call brief with account context, buyer hypotheses, recent signals, personas, questions, and talk tracks. +--- + +# Discovery Brief + +Prepare a discovery-call brief with account context, buyer hypotheses, recent signals, personas, questions, and talk tracks. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Identify account, attendee(s), product/motion, meeting objective, and known context. If attendee data is missing, still build an account-first brief.** +2. **Resolve and enrich the account with `search_companies` and `enrich_companies`.** +3. **Resolve attendees or target personas with `enrich_contacts` / `search_contacts`; capture title, function, seniority, and likely priorities.** +4. **Pull current context using `enrich_news`, `enrich_scoops`, and `enrich_intent`; keep only signals relevant to discovery.** +5. **Synthesize hypotheses, not claims. Label what is known vs inferred vs question-worthy.** +6. **Write discovery questions that tie directly to account facts, trigger signals, stakeholder role, and product relevance.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Short discovery brief with account snapshot, attendee notes, likely priorities, signal evidence, questions, and talk track. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/icp-definition/SKILL.md b/skills/icp-definition/SKILL.md new file mode 100644 index 0000000..5495b08 --- /dev/null +++ b/skills/icp-definition/SKILL.md @@ -0,0 +1,42 @@ +--- +name: icp-definition +description: Derive a reusable ICP from target accounts, winners, hypotheses, or market criteria using ZoomInfo firmographic and signal data. +--- + +# ICP Definition + +Derive a reusable ICP from target accounts, winners, hypotheses, or market criteria using ZoomInfo firmographic and signal data. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Capture the ICP purpose**: prospecting, campaign targeting, TAM sizing, scoring, territory planning, or product strategy. +2. **If seed accounts are provided, resolve and enrich them; if not, start from the user hypothesis and build candidate cohorts with `search_companies`.** +3. **Analyze common traits**: industry, size, revenue, growth, geography, tech stack, hierarchy, department footprint, intent, scoops, and news patterns. +4. **Distinguish hard filters from scoring signals; do not overfit to small seed lists.** +5. **Run broader/narrower searches to estimate addressability and identify filter cliffs.** +6. **Produce a reusable filter schema compatible with list building, TAM sizing, scoring, and campaign segmentation.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: ICP definition with hard filters, scoring signals, exclusions, addressability estimate, examples, and open assumptions. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/pipeline-hygiene/SKILL.md b/skills/pipeline-hygiene/SKILL.md new file mode 100644 index 0000000..4091a64 --- /dev/null +++ b/skills/pipeline-hygiene/SKILL.md @@ -0,0 +1,42 @@ +--- +name: pipeline-hygiene +description: Inspect open pipeline for stale deals, bad data, missing stakeholders, weak timing signals, and account-risk indicators. +--- + +# Pipeline Hygiene + +Inspect open pipeline for stale deals, bad data, missing stakeholders, weak timing signals, and account-risk indicators. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Identify the pipeline slice**: stage, owner, close period, segment, product, or named account list. +2. **For each account, resolve company identity and enrich firmographics, hierarchy, news, scoops, and intent.** +3. **Map active stakeholders using `search_contacts`, `get_recommended_contacts`, and buying-committee patterns; flag missing economic buyer, technical buyer, procurement, or executive sponsor.** +4. **Flag hygiene issues**: stale titles, departed contacts, duplicate accounts, low-confidence contact data, outdated firmographics, and weak next step. +5. **Score deal risk based on stakeholder coverage, recent signals, data freshness, account fit, and stage/close-date consistency.** +6. **Separate data fixes from deal-coaching recommendations so ops and sales can act independently.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Pipeline hygiene report with account/deal risks, data fixes, stakeholder gaps, signal evidence, and recommended action. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/procurement-prep/SKILL.md b/skills/procurement-prep/SKILL.md new file mode 100644 index 0000000..e4daa2d --- /dev/null +++ b/skills/procurement-prep/SKILL.md @@ -0,0 +1,42 @@ +--- +name: procurement-prep +description: Prepare for procurement by mapping stakeholders, account structure, risk signals, vendor/contract clues, and negotiation angles. +--- + +# Procurement Prep + +Prepare for procurement by mapping stakeholders, account structure, risk signals, vendor/contract clues, and negotiation angles. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify procurement context**: new purchase, expansion, renewal, security/legal review, pricing negotiation, or vendor consolidation. +2. **Resolve and enrich the account, including hierarchy, size, revenue, locations, and relevant technologies.** +3. **Search for procurement, finance, legal, security, IT, and executive stakeholders with accurate contact data.** +4. **Pull news and scoops for cost pressure, M&A, leadership changes, hiring, vendor consolidation, or compliance events.** +5. **Identify likely procurement concerns**: budget, risk, duplication, security, implementation effort, timing, and vendor leverage. +6. **Produce a prep plan with concessions to avoid, proof points to bring, and stakeholder-specific asks.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Procurement prep brief with stakeholder map, likely concerns, evidence, negotiation angles, and meeting checklist. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/renewal-risk-brief/SKILL.md b/skills/renewal-risk-brief/SKILL.md new file mode 100644 index 0000000..946b802 --- /dev/null +++ b/skills/renewal-risk-brief/SKILL.md @@ -0,0 +1,42 @@ +--- +name: renewal-risk-brief +description: Assess renewal or expansion risk using account signals, champion status, stakeholder gaps, growth/contraction clues, and next actions. +--- + +# Renewal Risk Brief + +Assess renewal or expansion risk using account signals, champion status, stakeholder gaps, growth/contraction clues, and next actions. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Capture customer/account, renewal date if known, product, current champions, sentiment, and open concerns.** +2. **Resolve and enrich the company, then pull relevant news, scoops, and intent for expansion, contraction, leadership change, M&A, or budget risk.** +3. **Verify champions with `enrich_contacts`; flag departures, title changes, low contact accuracy, and company moves.** +4. **Map renewal committee gaps across executive sponsor, business owner, technical owner, procurement, legal/security, and day-to-day users.** +5. **Score risk by champion health, signal trend, stakeholder coverage, company health, and data confidence.** +6. **Recommend retention actions and expansion plays separately.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Renewal risk brief with score, evidence, champion status, stakeholder gaps, retention actions, and expansion hypotheses. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/stakeholder-mapper/SKILL.md b/skills/stakeholder-mapper/SKILL.md new file mode 100644 index 0000000..8c67718 --- /dev/null +++ b/skills/stakeholder-mapper/SKILL.md @@ -0,0 +1,42 @@ +--- +name: stakeholder-mapper +description: Map stakeholders at a target account by role, seniority, likely influence, buying-committee coverage, and recommended outreach order. +--- + +# Stakeholder Mapper + +Map stakeholders at a target account by role, seniority, likely influence, buying-committee coverage, and recommended outreach order. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify the motion and account**: new logo, expansion, renewal, technical evaluation, procurement, executive alignment, or general mapping. +2. **Resolve the account through company ID, domain, or `search_companies` and enrich the company profile.** +3. **Use `lookup` for management levels, departments, and job functions; then search for contacts in priority persona groups.** +4. **Use `get_recommended_contacts` when the use case maps to prospecting, deal acceleration, or renewal/growth.** +5. **Classify contacts as economic buyer, technical buyer, evaluator, user champion, procurement/legal, executive sponsor, blocker, or unknown.** +6. **Rank outreach order by likely influence, relevance, seniority, contact accuracy, and coverage gaps.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Stakeholder map with roles, influence, evidence, contact confidence, gaps, and recommended outreach sequence. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/territory-planning/SKILL.md b/skills/territory-planning/SKILL.md new file mode 100644 index 0000000..3b89f0a --- /dev/null +++ b/skills/territory-planning/SKILL.md @@ -0,0 +1,42 @@ +--- +name: territory-planning +description: Design or rebalance territories using ZoomInfo account counts, fit, coverage, trigger density, and rep-actionable segments. +--- + +# Territory Planning + +Design or rebalance territories using ZoomInfo account counts, fit, coverage, trigger density, and rep-actionable segments. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Clarify planning objective**: carve new territory, rebalance, quota support, account assignment, industry/geography coverage, or white-space analysis. +2. **Translate territory rules into filters and resolve taxonomy values with `lookup`.** +3. **Use `search_companies` counts to size each territory and segment by geography, employee count, revenue, industry, and technology where relevant.** +4. **Pull sample accounts per territory and flag concentration risk, sparse data, or skew toward subsidiaries/parents.** +5. **Layer intent/scoops/news on representative or high-value accounts to estimate near-term actionability.** +6. **Recommend territory boundaries that balance account volume, ICP quality, revenue potential, and signal density.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Territory plan with counts, segment breakdowns, account samples, risks, and recommended assignment logic. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/trigger-based-list-build/SKILL.md b/skills/trigger-based-list-build/SKILL.md new file mode 100644 index 0000000..b7a48dd --- /dev/null +++ b/skills/trigger-based-list-build/SKILL.md @@ -0,0 +1,42 @@ +--- +name: trigger-based-list-build +description: Build account or contact lists around buying triggers such as intent, scoops, news, funding, hiring, technology, and leadership changes. +--- + +# Trigger-Based List Build + +Build account or contact lists around buying triggers such as intent, scoops, news, funding, hiring, technology, and leadership changes. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Define the trigger thesis**: what event indicates urgency, which personas matter, and what action the seller or marketer should take. +2. **Resolve company/contact filters with `lookup` and build the base universe with `search_companies` or `search_contacts`.** +3. **Layer trigger data**: use `enrich_intent` for topic demand, `enrich_scoops` for project/buying signals, and `enrich_news` for recent events. +4. **For contact lists, search or recommend contacts at triggered accounts and prioritize by persona fit and contact accuracy.** +5. **Rank records by trigger strength, ICP fit, recency, and activation readiness; separate high-priority, nurture, and review queues.** +6. **Include the evidence behind every trigger so the user can personalize outreach without hallucinated claims.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Prioritized triggered list with evidence, recency, recommended persona/action, confidence, and export-ready fields. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. diff --git a/skills/waterfall-enrich/SKILL.md b/skills/waterfall-enrich/SKILL.md new file mode 100644 index 0000000..7b4e750 --- /dev/null +++ b/skills/waterfall-enrich/SKILL.md @@ -0,0 +1,43 @@ +--- +name: waterfall-enrich +description: Enrich company and contact records with ZoomInfo-first matching, field-level confidence, provenance, unresolved rows, and activation-ready output. +--- + +# Waterfall Enrich + +Enrich company and contact records with ZoomInfo-first matching, field-level confidence, provenance, unresolved rows, and activation-ready output. + +## Input + +The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, contacts, domains, companies, CRM rows, segments, territories, product motion, timeframe, output format, or activation target. + +If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. + +## Workflow + +1. **Identify whether the input rows are companies, contacts, or mixed records; preserve the original row ID for auditability.** +2. **Resolve all requested fields and any filter values with `lookup` before search or enrichment; do not invent enum IDs.** +3. **For companies, resolve by domain first, then name using `search_companies`; enrich matched records with `enrich_companies`.** +4. **For contacts, resolve by email first, then name + company using `enrich_contacts`; fall back to `search_contacts` only when enrichment cannot identify a record.** +5. **Score each populated field as high / medium / low confidence based on exact match, ZoomInfo IDs, contact accuracy score, and recency signals.** +6. **Do not silently fill conflicting values. Route ambiguous, duplicate, low-confidence, and unmatched rows into review queues.** +7. **If the client exposes approved non-ZoomInfo enrichment providers, use them only for fields ZoomInfo cannot confidently fill and record source + cost. Otherwise mark the fallback as unavailable.** + +## Output Format + +Return a concise operator-ready artifact: + +- **Objective**: what this run is solving +- **Inputs used**: accounts, contacts, filters, timeframe, and assumptions +- **ZoomInfo workflow**: MCP tools used and why +- **Results**: Enriched table with original row ID, ZoomInfo IDs, populated fields, confidence, source/provenance, unresolved rows, conflicts, and recommended writeback action. +- **Evidence**: source fields, signal dates, confidence, and caveats +- **Next actions**: the 3-5 highest-leverage follow-ups + +## Quality Bar + +- Resolve taxonomy-backed fields with `lookup` before search or enrichment. +- Prefer stable identifiers: ZoomInfo IDs, company domains, and emails before fuzzy names. +- Keep confidence visible. Do not hide unmatched, ambiguous, stale, or conflicting records. +- Separate verified ZoomInfo data from inference and recommendations. +- Do not invent customer pain, internal initiatives, tools, budgets, or competitor usage without evidence. From 2e073c44364d84b4b9444d3a91b558df48a1821b Mon Sep 17 00:00:00 2001 From: Dominik Date: Thu, 21 May 2026 16:09:49 -0400 Subject: [PATCH 2/2] docs: align skills with GTM AI MCP tools --- skills/account-prioritization/SKILL.md | 12 ++++++++++++ skills/build-tam/SKILL.md | 11 +++++++++++ skills/business-case-builder/SKILL.md | 10 ++++++++++ skills/buying-committee/SKILL.md | 8 ++++---- skills/campaign-builder/SKILL.md | 13 ++++++++++++- skills/champion-tracking/SKILL.md | 11 +++++++++++ skills/clay-migration/SKILL.md | 11 +++++++++++ skills/competitive-battlecard/SKILL.md | 12 ++++++++++++ skills/crm-cleanup/SKILL.md | 10 ++++++++++ skills/deal-qualification/SKILL.md | 11 +++++++++++ skills/deal-risk-scorer/SKILL.md | 11 +++++++++++ skills/demo-scripting/SKILL.md | 10 ++++++++++ skills/discovery-brief/SKILL.md | 11 +++++++++++ skills/icp-definition/SKILL.md | 11 +++++++++++ skills/pipeline-hygiene/SKILL.md | 13 ++++++++++++- skills/procurement-prep/SKILL.md | 11 +++++++++++ skills/recommend-contacts/SKILL.md | 6 +++--- skills/renewal-risk-brief/SKILL.md | 10 ++++++++++ skills/stakeholder-mapper/SKILL.md | 13 ++++++++++++- skills/territory-planning/SKILL.md | 11 +++++++++++ skills/trigger-based-list-build/SKILL.md | 11 +++++++++++ skills/waterfall-enrich/SKILL.md | 11 +++++++++++ 22 files changed, 228 insertions(+), 10 deletions(-) diff --git a/skills/account-prioritization/SKILL.md b/skills/account-prioritization/SKILL.md index 878619d..4b63f03 100644 --- a/skills/account-prioritization/SKILL.md +++ b/skills/account-prioritization/SKILL.md @@ -13,6 +13,18 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use free discovery before credit-consuming depth: + +- `search_companies`: normalize or build the account universe. +- `search_intent`: discover accounts already researching relevant topics. +- `search_scoops`: discover accounts with recent events that create urgency. +- `enrich_companies`: deepen only the accounts being scored. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: add why-now evidence for known accounts. +- `find_recommended_contacts`: include reachability/stakeholder strength in the score when seller actionability matters. +- `account_research`: use for synthesized context when CRM/conversation history can change the ranking. + ## Workflow 1. **Capture the prioritization goal**: new logo, expansion, renewal, competitive takeout, territory focus, or campaign audience. diff --git a/skills/build-tam/SKILL.md b/skills/build-tam/SKILL.md index 0937a3a..dedc701 100644 --- a/skills/build-tam/SKILL.md +++ b/skills/build-tam/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use TAM tools by job type: + +- `lookup`: mandatory for industries, locations, technologies, job functions, and other controlled filters. +- `search_companies`: primary TAM counter and sample-account source; use `meta.totalResults` for counts. +- `find_similar_companies`: expand from seed/won accounts into lookalike ICP hypotheses before sizing. +- `search_intent`: size in-market subsegments when the TAM is topic-led or demand-led. +- `search_scoops`: size event-triggered subsegments such as funding, hiring, leadership changes, projects, or pain points. +- `enrich_companies`: enrich representative samples, not the whole TAM, unless the user explicitly approves credit-consuming enrichment. + ## Workflow 1. **Translate the user request into an explicit ICP**: industries, geography, employee count, revenue, technologies, growth/funding, ownership, and exclusions. diff --git a/skills/business-case-builder/SKILL.md b/skills/business-case-builder/SKILL.md index 5d352fd..c3ac012 100644 --- a/skills/business-case-builder/SKILL.md +++ b/skills/business-case-builder/SKILL.md @@ -13,6 +13,16 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use data tools to size and de-risk assumptions: + +- `account_research`: preferred when CRM, opportunity, or conversation context can inform the business case. +- `enrich_companies`: scale, revenue, employee footprint, technology, funding, and hierarchy assumptions. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: urgency and executive narrative. +- `find_recommended_contacts` and `search_contacts`: identify economic, technical, finance, and executive stakeholders. +- `contact_research`: use for named buyer-specific priorities when relationship history is relevant. + ## Workflow 1. **Clarify buyer persona, product, desired outcome, account, and any known metric assumptions.** diff --git a/skills/buying-committee/SKILL.md b/skills/buying-committee/SKILL.md index ff6f5f0..a3b72ec 100644 --- a/skills/buying-committee/SKILL.md +++ b/skills/buying-committee/SKILL.md @@ -28,7 +28,7 @@ Parallelize aggressively — once the company is resolved, account research, enr 1. **Anchor on purpose.** Read the research context from `$ARGUMENTS`. - If supplied, restate it in 1-2 sentences as the *map purpose* and keep it as the framing lens for every downstream step. - If missing, ask the user once for the context. If they decline or say "just general mapping", default to **general committee mapping for prospecting** and state that assumption at the top of the brief. - - From the context, derive: a **map purpose** (1 line), **priority personas/functions** (3-6 — e.g., CISO, SecOps Director, Procurement Lead, CFO), a **best-fit use case enum** for `get_recommended_contacts` (`PROSPECTING` / `DEAL_ACCELERATION` / `RENEWAL_AND_GROWTH`), and any **named hypotheses to test** (e.g., "find the actual economic buyer", "identify replacement champion"). These priorities drive the `account_research` query, `search_contacts` filters, scoops triage, and synthesis. + - From the context, derive: a **map purpose** (1 line), **priority personas/functions** (3-6 — e.g., CISO, SecOps Director, Procurement Lead, CFO), a **best-fit use case enum** for `find_recommended_contacts` (`PROSPECTING` / `DEAL_ACCELERATION` / `RENEWAL_AND_GROWTH`), and any **named hypotheses to test** (e.g., "find the actual economic buyer", "identify replacement champion"). These priorities drive the `account_research` query, `search_contacts` filters, scoops triage, and synthesis. 2. **Lookup metadata first** — call `lookup` for any fields relevant to the request (management levels, departments, job functions). Use returned `id` values in subsequent calls. @@ -39,11 +39,11 @@ Parallelize aggressively — once the company is resolved, account research, enr - Otherwise, call `search_companies` with the appropriate field (`companyWebsite` for a domain, `companyTicker` for a ticker, `companyName` for a name) and extract `companyId` from the top match. If no confident match, surface the ambiguity to the user before continuing rather than guessing. - Then `enrich_companies` for firmographics including `employeeCountByDepartment`. -5. **Fetch in parallel (retrieval, not filtering).** Treat each tool call as a context-retrieval step. Pull broadly now; decide what's relevant during synthesis. Steps that only need the `companyId` (plus inputs from step 1) can run in parallel — `account_research`, `get_recommended_contacts`, `search_contacts`, `enrich_scoops`/`search_scoops`. +5. **Fetch in parallel (retrieval, not filtering).** Treat each tool call as a context-retrieval step. Pull broadly now; decide what's relevant during synthesis. Steps that only need the `companyId` (plus inputs from step 1) can run in parallel — `account_research`, `find_recommended_contacts`, `search_contacts`, `enrich_scoops`/`search_scoops`. - **Tailor the `account_research` query to the map purpose.** Don't pass a generic "tell me about this account" string. Inject the full research context — name the deal stage, the offering, named hypotheses, persona priorities — and ask for named individuals organized by function, engagement status, deal context (stage, last activity, competition), and any signals about budget owners, blockers, or champions. The more context the better. - Normalize engagement to two states: **Engaged** (explicit prior interaction signal) or **New** (everything else, including ambiguous). Default to New when in doubt. - If `account_research` surfaces dates in the past (renewal, contract end, last activity), retain them but tag for verification — they may signal active negotiation, broken CRM sync, or a missed milestone. - - **`get_recommended_contacts`** — pass the use-case enum derived in step 1. Treat as supplemental signal. Empty results are common (cold-start tenants, no CRM data); note as a confidence indicator rather than retrying. + - **`find_recommended_contacts`** — pass the use-case enum derived in step 1. Treat as supplemental signal. Empty results are common (cold-start tenants, no CRM data); note as a confidence indicator rather than retrying. - **`search_contacts`** — filter by the priority personas/functions derived in step 1 (resolved against `lookup` IDs from step 2 and GTM personas from step 3). Sort by `-contactAccuracyScore`. Pull broader than you'll keep — filtering happens in step 7. - **`enrich_scoops` / `search_scoops`** — 90-day window, no role filter. Retrieval is unconstrained; step 7 will triage for VP+ moves relevant to the priority personas and named hypotheses. @@ -53,7 +53,7 @@ Parallelize aggressively — once the company is resolved, account research, enr 7. **Synthesize.** Each retrieval is raw context — now decide what makes the map, framed by the map purpose, priority personas, and named hypotheses. Apply these principles: - **Scoops triage**: review every scoop returned. Keep **New Hire / Lateral Move / Executive Move / Promotion** at VP+ level, especially in the priority personas or adjacent functions named in the context. For each newly-named person, run `search_contacts` if not already enriched and add to the committee with a `RECENTLY APPOINTED` flag and the event date. These often pre-date what `account_research` knows. Drop scoops irrelevant to the map purpose. - - **Search/recommendation triage**: from the broad `search_contacts` and `get_recommended_contacts` pulls, keep contacts that fit a priority persona, address a named hypothesis (e.g., "find the actual economic buyer"), or fill a coverage gap visible from `account_research`. Drop everyone else — broad retrieval is fine, broad output isn't. + - **Search/recommendation triage**: from the broad `search_contacts` and `find_recommended_contacts` pulls, keep contacts that fit a priority persona, address a named hypothesis (e.g., "find the actual economic buyer"), or fill a coverage gap visible from `account_research`. Drop everyone else — broad retrieval is fine, broad output isn't. - **Role classification (conservative)**: - **Champions** require explicit engagement evidence (CRM activity, demo attended, prior emails). Title alone is never sufficient — without a signal, place under Influencers > Potential Champions. - **Technical Evaluators** are Director+ in IT, Engineering, or the function being sold to. Sales Ops VPs are Influencers > Operations unless the product evaluates against criteria they own. diff --git a/skills/campaign-builder/SKILL.md b/skills/campaign-builder/SKILL.md index f77138a..6446dc8 100644 --- a/skills/campaign-builder/SKILL.md +++ b/skills/campaign-builder/SKILL.md @@ -13,11 +13,22 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Select tools based on campaign source: + +- ICP/firmographic campaign: `lookup` → `search_companies` → optional `enrich_companies`. +- Persona campaign: `search_contacts` after account or company filters are resolved. +- Intent campaign: `search_intent` for net-new in-market accounts; `enrich_intent` for known-account depth. +- Event campaign: `search_scoops` for trigger discovery; `enrich_scoops` and `enrich_news` for evidence on selected accounts. +- Account-based stakeholder campaign: `find_recommended_contacts`; add `contact_research` only when relationship history matters. +- Lookalike campaign: `find_similar_companies` or `find_similar_contacts` from seed wins/champions. + ## Workflow 1. **Clarify campaign goal, offer, target segment, exclusion rules, and success metric. If absent, assume outbound prospecting and state it.** 2. **Build the account universe with `search_companies`; resolve filters through `lookup` first.** -3. **Identify target personas using `search_contacts`, `get_recommended_contacts`, or buying-committee logic depending on whether the account list is known.** +3. **Identify target personas using `search_contacts`, `find_recommended_contacts`, or buying-committee logic depending on whether the account list is known.** 4. **Enrich account context with `enrich_companies`, `enrich_intent`, `enrich_scoops`, and `enrich_news` for segmentation and proof points.** 5. **Create practical segments**: fit tier, trigger type, persona, industry, company size, and activation priority. 6. **Draft campaign angles that cite observed data only; never invent initiatives, tools, or pains not supported by evidence.** diff --git a/skills/champion-tracking/SKILL.md b/skills/champion-tracking/SKILL.md index db948bd..8218f5d 100644 --- a/skills/champion-tracking/SKILL.md +++ b/skills/champion-tracking/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use contact-first tools: + +- `enrich_contacts`: verify current employment, title, seniority, contact details, and ZoomInfo contact ID. +- `contact_research`: use for high-value champions when CRM/conversation history can explain relationship strength. +- `find_similar_contacts`: find champion lookalikes when the original champion profile is known. +- `search_contacts`: find replacement champions at the old account or equivalent contacts at the new account. +- `find_recommended_contacts`: identify who else to engage at the original account after champion movement. +- `enrich_companies`: qualify the champion's new company for expansion or new-logo opportunity. + ## Workflow 1. **Identify known champions by email, name, company, CRM/contact ID, or supplied list. Preserve original account association.** diff --git a/skills/clay-migration/SKILL.md b/skills/clay-migration/SKILL.md index 95e5967..456c91f 100644 --- a/skills/clay-migration/SKILL.md +++ b/skills/clay-migration/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Map Clay steps to current GTM AI MCP tools: + +- Discovery rows: `search_companies`, `search_contacts`, `search_intent`, and `search_scoops` depending on whether the table starts from ICP, persona, intent topic, or event trigger. +- Lookalikes: `find_similar_companies` for account expansion and `find_similar_contacts` for persona replication from a known champion/buyer. +- Enrichment: `enrich_companies` and `enrich_contacts`, in batches up to 25 after free search resolution. +- Why-now context: `enrich_intent`, `enrich_scoops`, and `enrich_news` for known companies. +- Stakeholder selection: `find_recommended_contacts` for account-level who-to-contact decisions. +- Narrative research: `account_research` and `contact_research` when the Clay workflow previously depended on CRM or conversation history. + ## Workflow 1. **Ask for or infer the current Clay table purpose**: account list, contact list, enrichment, scoring, personalization, or campaign prep. diff --git a/skills/competitive-battlecard/SKILL.md b/skills/competitive-battlecard/SKILL.md index 1fc210e..e9bf06b 100644 --- a/skills/competitive-battlecard/SKILL.md +++ b/skills/competitive-battlecard/SKILL.md @@ -13,6 +13,18 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use direct research plus discovery tools: + +- `account_research`: account/deal-specific battlecards when first-party context matters. +- `search_companies` / `enrich_companies`: resolve and profile competitors and target accounts. +- `search_intent`: discover accounts researching competitor or category topics. +- `enrich_intent`: inspect known account or competitor intent topics. +- `search_scoops` / `enrich_scoops`: event-led competitive triggers. +- `enrich_news`: public narrative, launches, funding, M&A, and executive moves. +- `find_similar_companies`: identify adjacent competitors or lookalike displacement targets. + ## Workflow 1. **Clarify whether the battlecard is competitor-led, account-led, or deal-led, and what decision it should support.** diff --git a/skills/crm-cleanup/SKILL.md b/skills/crm-cleanup/SKILL.md index fcdee19..78820c4 100644 --- a/skills/crm-cleanup/SKILL.md +++ b/skills/crm-cleanup/SKILL.md @@ -13,6 +13,16 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Keep cleanup deterministic and credit-aware: + +- `lookup`: normalize controlled CRM values before comparing or writing. +- `search_companies` / `search_contacts`: free matching pass before enrichment. +- `enrich_companies` / `enrich_contacts`: source-of-truth values after a confident match; batch up to 25 records per call. +- `find_recommended_contacts`: not for field cleanup; use it only if cleanup expands into account coverage recommendations. +- `account_research` / `contact_research`: avoid for bulk cleanup; use only for high-value ambiguous records requiring relationship context. + ## Workflow 1. **Classify the cleanup request**: stale fields, missing fields, duplicates, bad domains, inactive contacts, title drift, account hierarchy, or territory/segment hygiene. diff --git a/skills/deal-qualification/SKILL.md b/skills/deal-qualification/SKILL.md index 8f9a38b..1f7f528 100644 --- a/skills/deal-qualification/SKILL.md +++ b/skills/deal-qualification/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use qualification-specific evidence: + +- `account_research`: preferred when the decision depends on CRM, opportunity, or conversation history. +- `enrich_companies`: firmographic, technographic, hierarchy, and growth fit. +- `enrich_intent`: known-account buyer-intent topics and trend. +- `enrich_scoops` / `enrich_news`: event urgency and public narrative. +- `find_recommended_contacts`: stakeholder/actionability strength. +- `search_contacts`: explicit persona coverage gaps. + ## Workflow 1. **Clarify qualification framework if supplied; otherwise use fit, pain/trigger, stakeholder access, timing, and data confidence.** diff --git a/skills/deal-risk-scorer/SKILL.md b/skills/deal-risk-scorer/SKILL.md index 01e06df..a4211a3 100644 --- a/skills/deal-risk-scorer/SKILL.md +++ b/skills/deal-risk-scorer/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use the account-specific enrichment stack: + +- `account_research`: preferred when the user asks for deal/account risk and first-party CRM or conversation context may exist. +- `enrich_companies`: baseline company, hierarchy, firmographic, and technology context. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: timing, urgency, and disruption evidence for the known account. +- `find_recommended_contacts`: stakeholder coverage and deal-acceleration contact recommendations. +- `search_contacts`: targeted gaps for economic buyer, technical buyer, procurement, finance, legal, or executive sponsor. +- `contact_research`: named stakeholder relationship/role context. + ## Workflow 1. **Start from named opportunity/account context and clarify deal stage, close date, product, known stakeholders, and primary concern if available.** diff --git a/skills/demo-scripting/SKILL.md b/skills/demo-scripting/SKILL.md index bb73340..9d5a77f 100644 --- a/skills/demo-scripting/SKILL.md +++ b/skills/demo-scripting/SKILL.md @@ -13,6 +13,16 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use tools to ground demo narrative in evidence: + +- `account_research`: synthesized account context and first-party relationship history when available. +- `contact_research`: attendee-specific background and prior interaction context. +- `enrich_companies` / `enrich_contacts`: structured facts about company and attendees. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: current why-now hooks and proof points. +- `find_recommended_contacts`: identify missing demo stakeholders or follow-up audience. + ## Workflow 1. **Capture product/demo objective, account, attendees/personas, and desired outcome.** diff --git a/skills/discovery-brief/SKILL.md b/skills/discovery-brief/SKILL.md index afdf7f7..98c1137 100644 --- a/skills/discovery-brief/SKILL.md +++ b/skills/discovery-brief/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Blend direct data with context agents: + +- `account_research`: preferred for account briefing when CRM/conversation context may exist. +- `contact_research`: use for named attendees when relationship history matters. +- `enrich_companies` / `enrich_contacts`: pull structured account and attendee facts. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: known-account why-now evidence. +- `find_recommended_contacts`: if attendee list is incomplete and the seller needs who else to involve. +- `search_contacts`: fill missing persona gaps after recommendations. + ## Workflow 1. **Identify account, attendee(s), product/motion, meeting objective, and known context. If attendee data is missing, still build an account-first brief.** diff --git a/skills/icp-definition/SKILL.md b/skills/icp-definition/SKILL.md index 5495b08..6d77cb8 100644 --- a/skills/icp-definition/SKILL.md +++ b/skills/icp-definition/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Build the ICP from both fit and timing signals: + +- `enrich_companies`: profile seed/won accounts. +- `find_similar_companies`: generate lookalike hypotheses from best-fit customers. +- `search_companies`: test hard filters and estimate addressability. +- `search_intent`: identify topic-led in-market traits and demand clusters. +- `search_scoops`: identify event triggers correlated with urgency. +- `account_research`: use when first-party CRM/conversation history can explain why accounts won or stalled. + ## Workflow 1. **Capture the ICP purpose**: prospecting, campaign targeting, TAM sizing, scoring, territory planning, or product strategy. diff --git a/skills/pipeline-hygiene/SKILL.md b/skills/pipeline-hygiene/SKILL.md index 4091a64..6885ba8 100644 --- a/skills/pipeline-hygiene/SKILL.md +++ b/skills/pipeline-hygiene/SKILL.md @@ -13,11 +13,22 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Prioritize tools that reveal deal risk and bad data: + +- `enrich_companies` and `enrich_contacts`: verify account/contact freshness, firmographics, titles, and reachability. +- `find_recommended_contacts`: identify missing stakeholders and stronger contacts for deal acceleration. +- `contact_research`: inspect relationship history for named stakeholders when conversation data matters. +- `account_research`: synthesize CRM, opportunity, and account context when available. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: add known-account risk and urgency signals. +- `search_contacts`: fill specific committee gaps when recommended contacts are insufficient. + ## Workflow 1. **Identify the pipeline slice**: stage, owner, close period, segment, product, or named account list. 2. **For each account, resolve company identity and enrich firmographics, hierarchy, news, scoops, and intent.** -3. **Map active stakeholders using `search_contacts`, `get_recommended_contacts`, and buying-committee patterns; flag missing economic buyer, technical buyer, procurement, or executive sponsor.** +3. **Map active stakeholders using `search_contacts`, `find_recommended_contacts`, and buying-committee patterns; flag missing economic buyer, technical buyer, procurement, or executive sponsor.** 4. **Flag hygiene issues**: stale titles, departed contacts, duplicate accounts, low-confidence contact data, outdated firmographics, and weak next step. 5. **Score deal risk based on stakeholder coverage, recent signals, data freshness, account fit, and stage/close-date consistency.** 6. **Separate data fixes from deal-coaching recommendations so ops and sales can act independently.** diff --git a/skills/procurement-prep/SKILL.md b/skills/procurement-prep/SKILL.md index e4daa2d..3891d56 100644 --- a/skills/procurement-prep/SKILL.md +++ b/skills/procurement-prep/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use tools that expose committee and negotiation context: + +- `account_research`: preferred for procurement prep when CRM/conversation history may reveal deal status, objections, and relationship context. +- `enrich_companies`: structure, size, hierarchy, technologies, and financial context. +- `find_recommended_contacts`: recommended procurement-adjacent and executive contacts. +- `search_contacts`: explicit procurement, legal, finance, security, IT, and executive searches. +- `enrich_scoops` / `enrich_news`: layoffs, funding, M&A, leadership changes, vendor consolidation, or cost pressure. +- `enrich_intent`: known-account evaluation topics. + ## Workflow 1. **Clarify procurement context**: new purchase, expansion, renewal, security/legal review, pricing negotiation, or vendor consolidation. diff --git a/skills/recommend-contacts/SKILL.md b/skills/recommend-contacts/SKILL.md index 83a3a1a..b78d880 100644 --- a/skills/recommend-contacts/SKILL.md +++ b/skills/recommend-contacts/SKILL.md @@ -1,5 +1,5 @@ --- -name: recommended-contacts +name: recommend-contacts description: Get AI-powered contact recommendations at a target company based on your ZoomInfo interaction history. Provide a company name or domain and optionally a use case. --- @@ -29,7 +29,7 @@ The user will provide via `$ARGUMENTS`: - "deal acceleration" or "new business" → `DEAL_ACCELERATION` (based on contacts in closed-won CRM opportunities for new business) - "renewal", "growth", or "expansion" → `RENEWAL_AND_GROWTH` (based on contacts in closed-won CRM opportunities for renewals) -5. **Get recommendations** using `get_recommended_contacts` with: +5. **Get recommendations** using `find_recommended_contacts` with: - `ziCompanyId`: the resolved ZoomInfo company ID - `useCaseType`: the mapped enum value - `pageSize`: user-specified count or 25 @@ -74,7 +74,7 @@ Rank the top 5 contacts to engage first, with reasoning: ### Next Steps - Use `/zoominfo:enrich-contact` to deep-dive on any specific person -- Use `/zoominfo:find-buyers` if you need to filter by specific persona criteria beyond what recommendations provide +- Use `/zoominfo:recommend-contacts` if you need to filter by specific persona criteria beyond what recommendations provide - If recommendations are sparse, note that PROSPECTING recommendations improve as you use ZoomInfo more (view, copy, export contacts). DEAL_ACCELERATION and RENEWAL_AND_GROWTH require CRM integration. ### Important Notes on Scores diff --git a/skills/renewal-risk-brief/SKILL.md b/skills/renewal-risk-brief/SKILL.md index 946b802..5b77a32 100644 --- a/skills/renewal-risk-brief/SKILL.md +++ b/skills/renewal-risk-brief/SKILL.md @@ -13,6 +13,16 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use renewal-specific account and contact context: + +- `account_research`: preferred for renewal risk because CRM and conversation history often determine true risk. +- `contact_research`: inspect named champions, detractors, and executive sponsors. +- `enrich_contacts`: verify champion movement, title changes, reachability, and employment status. +- `find_recommended_contacts`: renewal/growth use case for missing stakeholders and expansion paths. +- `enrich_companies`, `enrich_intent`, `enrich_scoops`, `enrich_news`: company health, signal trend, and disruption evidence. + ## Workflow 1. **Capture customer/account, renewal date if known, product, current champions, sentiment, and open concerns.** diff --git a/skills/stakeholder-mapper/SKILL.md b/skills/stakeholder-mapper/SKILL.md index 8c67718..97461f1 100644 --- a/skills/stakeholder-mapper/SKILL.md +++ b/skills/stakeholder-mapper/SKILL.md @@ -13,12 +13,23 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use the contact stack in order: + +- `find_recommended_contacts`: first choice for who-to-contact recommendations at a named account because it uses ZoomInfo activity and CRM win patterns when available. +- `search_contacts`: fill explicit persona or committee gaps not covered by recommendations. +- `enrich_contacts`: reveal full contact data only for selected contacts, in batches up to 25. +- `contact_research`: synthesize CRM/conversation context for named high-priority stakeholders. +- `account_research`: use when the stakeholder map depends on account history, deal stage, or relationship context. +- `lookup`: resolve management level, department, function, and location filters before searching. + ## Workflow 1. **Clarify the motion and account**: new logo, expansion, renewal, technical evaluation, procurement, executive alignment, or general mapping. 2. **Resolve the account through company ID, domain, or `search_companies` and enrich the company profile.** 3. **Use `lookup` for management levels, departments, and job functions; then search for contacts in priority persona groups.** -4. **Use `get_recommended_contacts` when the use case maps to prospecting, deal acceleration, or renewal/growth.** +4. **Use `find_recommended_contacts` when the use case maps to prospecting, deal acceleration, or renewal/growth.** 5. **Classify contacts as economic buyer, technical buyer, evaluator, user champion, procurement/legal, executive sponsor, blocker, or unknown.** 6. **Rank outreach order by likely influence, relevance, seniority, contact accuracy, and coverage gaps.** diff --git a/skills/territory-planning/SKILL.md b/skills/territory-planning/SKILL.md index 3b89f0a..716aea4 100644 --- a/skills/territory-planning/SKILL.md +++ b/skills/territory-planning/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use tools for both capacity balance and actionability: + +- `lookup`: normalize regions, industries, technologies, and size bands. +- `search_companies`: count and sample accounts by proposed territory rules. +- `search_intent`: compare in-market account density across territories. +- `search_scoops`: compare event-trigger density across territories. +- `find_similar_companies`: expand seed customer profiles into territory-specific lookalike pools. +- `enrich_companies`: enrich representative accounts only, unless deeper credit-consuming analysis is approved. + ## Workflow 1. **Clarify planning objective**: carve new territory, rebalance, quota support, account assignment, industry/geography coverage, or white-space analysis. diff --git a/skills/trigger-based-list-build/SKILL.md b/skills/trigger-based-list-build/SKILL.md index b7a48dd..d34da0c 100644 --- a/skills/trigger-based-list-build/SKILL.md +++ b/skills/trigger-based-list-build/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Start from the trigger type: + +- Topic/keyword demand: `search_intent` to discover companies researching a topic. +- Business event: `search_scoops` to discover companies with recent leadership, funding, hiring, layoffs, project, pain-point, expansion, or launch events. +- Known account deepening: `enrich_intent`, `enrich_scoops`, and `enrich_news`. +- Persona activation: `search_contacts` for explicit roles or `find_recommended_contacts` for account-specific who-to-contact decisions. +- Lookalike trigger expansion: `find_similar_companies` or `find_similar_contacts` from known good accounts/champions. +- Full details: `enrich_companies` and `enrich_contacts` only for selected records. + ## Workflow 1. **Define the trigger thesis**: what event indicates urgency, which personas matter, and what action the seller or marketer should take. diff --git a/skills/waterfall-enrich/SKILL.md b/skills/waterfall-enrich/SKILL.md index 7b4e750..fa18ad0 100644 --- a/skills/waterfall-enrich/SKILL.md +++ b/skills/waterfall-enrich/SKILL.md @@ -13,6 +13,17 @@ The user will describe the job via `$ARGUMENTS`. Capture any supplied accounts, If a required entity is missing, ask one concise clarifying question. If the request is broad but actionable, proceed with a stated assumption rather than blocking. +## MCP Tool Selection + +Use the full MCP stack deliberately: + +- `lookup`: resolve every taxonomy-backed field before search or enrichment. +- `search_companies` / `search_contacts`: free pre-resolution before spending bulk credits. +- `enrich_companies` / `enrich_contacts`: batch up to 25 records per call only after entity match confidence is acceptable. +- `search_intent` / `search_scoops`: use only when the enrichment run should discover net-new high-signal accounts, not for simple field completion. +- `enrich_intent`, `enrich_scoops`, `enrich_news`: use for known companies when activation needs why-now evidence. +- `account_research` / `contact_research`: use for synthesized CRM/conversation context only when the output requires a narrative brief, not raw field fill. + ## Workflow 1. **Identify whether the input rows are companies, contacts, or mixed records; preserve the original row ID for auditability.**