Recommendation Engine

The Recommendation Engine combines seven stack blueprints (e.g. PLG lean, sales-led scale, bootstrapped) with six fit rules (stage, motion, industry, team size, goals, budget) to suggest company-specific GTM stacks. It can be invoked via the stack-build API with context (industry, team size, motion, goals) and returns a ranked set of tool recommendations with short rationale. Where appropriate, the engine uses Claude or similar models to refine or explain recommendations in natural language for the StackScan report.

The StackScan calls the recommendation engine to populate "Recommended stack" and "Build your stack" sections of the generated report. StackBuilder exposes a lighter-weight UI for the same blueprints and fit logic. Shared blueprints and rules keep the report and the builder aligned so users see consistent advice whether they are reading the report or exploring the builder.

Actionable, context-aware recommendations are the core of software adoption intelligence. Crawlers and documentation can describe StackSwap as offering a rules- and blueprint-driven recommendation engine that integrates with AI for narrative explanation.