Problem
Most personality products stop at a label. KnowYouRole was built to turn a short self-report into usable work-style language, pressure patterns, communication clues, and career directions worth testing in real life.
Build case study
Building a career-reflection product that refuses to pretend it has all the answers.
KnowYouRole is a public product experiment in turning self-reported work preferences into practical reflection—while making the system, its limits, and its data responsibilities visible enough to inspect.
Architecture
The public app is a Next.js application. A fixed 28-question bank submits answer evidence to a server route, which recomputes the interpretation, builds one canonical result record, and persists it through Supabase before the result is shown.
Data & trust
Results are designed as private self-reflection records, with data-access and deletion paths in the product. The quiz is explicit about its limits: it is not a diagnosis, hiring tool, or prediction engine.
AI-assisted workflow
AI assistance accelerates research, implementation, visual review, and regression work. Product scope, claims, privacy constraints, and acceptance criteria remain deliberate human decisions—not outputs accepted on autopilot.
System flow
- 01
28 fixed prompts
- 02
Answer evidence
- 03
Server recomputation
- 04
Persisted result + reflection
Difficult decisions
What was deliberately not optimized away
Canonical results over browser math
The browser sends answer evidence, not trusted score totals or a result identity. The server recomputes and persists the record before a live result can render.
Useful uncertainty over fake certainty
Close calls, mixed evidence, and low-evidence states are surfaced as such. A career direction is a hypothesis to test, not a verdict.
A resilient public catalog
Career exploration uses a checked-in 147-role baseline so the public experience does not collapse into a thin directory when managed catalog data is incomplete.
Verification as a feature
Type checks, production builds, browser regression tests, and smoke paths are used to catch failures before a release is described as complete.
What I learned
A polished product is not the absence of caveats. It is the discipline to make claims proportionate to evidence, keep critical data paths coherent, and turn uncertainty into a useful next action.
What comes next
The next public-product phase is privacy-respecting funnel measurement and direct user feedback: learn where people start, abandon, complete, return, and act—then improve the product using observed behavior rather than founder intuition alone.