A portfolio project built to show practical AI-assisted product engineering
Business Launch AI is a Tennessee-focused opportunity analyzer that turns structured market signals, business constraints, and user preferences into a clear small-business report. It is built as a deterministic product demo: the recommendations come from scoring logic, evidence rules, validation checks, and explanation layers rather than an unbounded live AI chat response.
What this project demonstrates
Each major product decision reflects a deliberate engineering or design choice.
Structured recommendation logic
Opportunities are scored, ranked, and explained using a deterministic stack of opportunity profiles, market weights, score breakdowns, and reason codes — not a prompt.
Founder vs. owner personalization
Two distinct paths with different field sets, section flows, report sections, and output language depending on whether the user is starting fresh or expanding an existing business.
Cost-to-first-paid-job framing
A budget question scoped to the first paid service job rather than a full startup, which narrows the decision and keeps the estimate actionable for low-budget users.
Tennessee geography handling
Location inputs (city, ZIP, county) resolve to a structured Tennessee market profile that shapes the scoring, evidence, and support-match layers throughout the report.
Market evidence and explanation layers
Reports surface scored reasons, near misses, market signal chips, and a tradeoff note so the user can see why the top match won, not just what it is.
What-if recommendation refiner
An interactive refiner re-ranks candidates against time, work style, goal, and risk axes and shows exactly what moved up, what moved down, and why.
Validation-heavy development workflow
Validator scripts, audit scripts, and narrative safety checks run against the catalog to catch scoring drift, copy regressions, and internal-language leaks before commits.
AI-assisted product engineering
Built with AI assistance throughout the design and development process — as a tool for accelerating structured product work, not as a live LLM feature presented to the user.
How the intelligence works
The app is deterministic and reproducible. The same inputs always produce the same report.
Scoring and ranking
Opportunity profiles are scored against market traits, budget fit, founder style, and eligibility signals. Score breakdowns and factor labels explain each result.
Evidence and explanation
Market evidence bridges connect scored signals to readable reasons. Near misses, tradeoffs, and the advisor note expose the logic without hiding it behind a black box.
Safety and validation
Narrative safety rules, audit scripts, and copy validators guard against guarantee language, internal-string leaks, and scoring drift across the full opportunity catalog.
This is an independent portfolio project. It is not affiliated with the State of Tennessee, any state agency, or any official business-support program. Reports are informational only and do not constitute financial, legal, or business advice. Always verify funding, support eligibility, and market conditions directly with official sources.