AI-assisted RAID analysis and stakeholder-tiered project status reporting.
agentic-pm converts raw task data into structured RAID analysis, workstream RAG scoring, delivery predictability scores, and stakeholder-ready status reports — reducing weekly reporting from hours to seconds.
Built by a delivery manager, for delivery managers. Grounded in real programme delivery across ERP, cloud infrastructure, telecoms, and AI SaaS environments.
Every delivery manager maintains a RAID log. Most spend 2-4 hours every week turning that log into a status report, a steering committee summary, and a set of escalations — manually, from scratch, in a format that has to change depending on who is reading it.
agentic-pm automates that conversion. You maintain one task list. It produces the right report for the right audience in seconds.
- Reads your task data (CSV, JSON, or pasted text)
- Scores each task as HIGH / MEDIUM / LOW risk based on status, blockers, and due dates
- Groups tasks by workstream and assigns an overall RAG (RED / AMBER / GREEN) per workstream
- Identifies Risks, Actions, Issues, and Decisions from the task data automatically
- Calculates a delivery predictability score (0-100)
- Produces a stakeholder-tiered status report in three formats:
- Delivery — full task-level detail, all blockers, recommended actions
- Executive — one paragraph per workstream, key risks and decisions only
- Board — three sentences, overall RAG, single biggest risk, single decision required
The 82% to 95% delivery predictability improvement documented across regulated infrastructure programmes came from consistent, structured status reporting — not heroics. The discipline of weekly RAID analysis and stakeholder-appropriate communication is what keeps programmes on track.
agentic-pm makes that discipline repeatable and low-effort, regardless of programme size or team experience.
pip install -r requirements.txt
export ANTHROPIC_API_KEY=your-key-here
# Generate sample data and run
python agentic_pm.py --sample
python agentic_pm.py --input sample_tasks.csv
# Executive-level report
python agentic_pm.py --input sample_tasks.csv --stakeholder executive
# Board-level report
python agentic_pm.py --input sample_tasks.csv --stakeholder boardA hosted web version is available at agentic-pm.onrender.com
Features:
- CSV upload or manual task entry
- Workstream RAG dashboard
- Export as Word (.doc), plain text (.txt), or PDF
- No installation required
CSV with columns: Workstream, Task name, Owner, Due date, Status, Blockers
Or pipe-separated text: Workstream | Task name | Owner | Due date | Status | Blockers
Status values: Complete | In Progress | At Risk | Not Started
| Condition | Risk level |
|---|---|
| Status = Complete | LOW |
| Status = At Risk | HIGH |
| Any blocker present (not None/N/A) | HIGH |
| Due date within 7 days and not complete | HIGH |
| Due date within 14 days | MEDIUM |
| Status = Not Started with due date within 21 days | MEDIUM |
| Everything else | LOW |
Workstream RAG = RED if any HIGH risk task exists / AMBER if any MEDIUM / GREEN if all LOW.
- Workstream RAG summary per workstream
- Full RAID-structured weekly status report
- Delivery predictability score with calculation breakdown
- Steering committee summary paragraph
- Escalation and decisions required section
- Python 3.9+
- Anthropic API key (set as
ANTHROPIC_API_KEYenvironment variable)
- Azure DevOps native export parser
- Jira CSV export parser
- Slack / Teams notification integration
- Power BI dashboard export
- Multi-workstream dependency mapping
- Historical predictability trend tracking
Pull requests welcome. Please open an issue first to discuss what you would like to change.
MIT — see LICENSE
Part of an open-source toolkit for AI-assisted project delivery. See also: cutover-copilot | erp-discovery-agent | prince2-agile-templates