Skip to content

Latest commit

 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

VisionCaster

Validate the problem before you build the solution.

A PM toolkit for AI-powered problem validation. Give it a problem statement, it does the research, surfaces what's actually known, and generates a structured, evidence-backed problem brief — before you commit to building anything.

Built in October 2025. MVP works end-to-end. Paused — may become part of a larger PM specialist toolkit.


The core tool: ProblemExplorer

Most teams commit to building something before they've truly validated the problem. ProblemExplorer flips that — AI does the research first, so you're working from evidence, not assumptions.

make problemexplorer PROBLEM="Managing AI-first teams is challenging"

What it does:

  1. Takes your problem statement
  2. Generates research queries automatically
  3. Executes searches and synthesizes findings
  4. Validates evidence quality
  5. Outputs a structured, research-backed problem statement document

Why this matters for PMs: A problem statement you can defend — with sources, evidence dimensions, and quality validation — is worth 10 built-on-assumptions ones.


What's inside

Tool Status Description
ProblemExplorer ✅ MVP complete AI-research-first problem validation
ProblemSpace v1 📦 Archived First attempt — solved the wrong problem

On ProblemSpace v1: The first version asked users to provide evidence and had AI format it. User testing revealed the flaw immediately — people haven't done the research yet, that's the whole point. "People are lazy — AI should do the work." That insight became ProblemExplorer. ProblemSpace v1 is kept as a learning artifact.


Setup

Prerequisites: Python 3.9+, Amplifier (github.com/microsoft/amplifier)

git clone https://github.com/cpark4x/visioncaster
cd visioncaster
pip install -e amplifier/.

Note: ProblemExplorer MVP uses mock search data. Production use requires a real web search integration (SerpAPI or similar).


Key learnings from building this

  1. MVP = working end-to-end — mock data is fine, complete workflow matters
  2. Test the assumption, not the polish — v1 looked good and solved the wrong problem entirely
  3. User insight beats architecture — one round of testing revealed more than weeks of design
  4. Code for structure, AI for intelligence — reliable pipeline + intelligent AI > trying to make AI reliable

Status

Working MVP. Paused. Will return when integrating into a broader PM specialist toolkit — likely alongside Outcomist and canvas-specialists research pipeline.


Built by

Chris Park — Senior PM, Microsoft Office of the CTO, AI Incubation group. Engineering degree from Waterloo. 17 years shipping product.

LinkedIn · GitHub

About

Visioncaster project

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages