Synthetic Market Research

What it does

This skill automates directional market research by generating persona‑conditioned free‑text survey responses with an LLM and converting those responses into Likert-like probability ratings using Semantic Similarity Rating (SSR). It produces PMF distributions, mean scores, segment breakdowns, qualitative themes, and a methodology summary saved as a markdown report.

When to use it

Use when you need quick, inexpensive validation of product concepts, pricing, or purchase intent without recruiting real panels — for early-stage founders, product managers, and researchers who want directional signals before running real user studies. Also useful for A/B comparisons of concepts or pricing tiers.

What's included

• Scripts: example Python SSR pipeline (references present in the skill directory)
• References: SSR_METHODOLOGY.md and example runs for concept and pricing tests
• Instructions: persona generation, prompt templates, SSR reference sets, recommended sampling (2 responses/persona), and how to save results to output/research_<concept>_<timestamp>.md

Compatible agents

Works with agents that can call LLM APIs and run Python/batch scripts (OpenAI/Anthropic-backed agents, Claude, Copilot-style automation). It expects access to an LLM and the semantic-similarity-rating package.