FairPost — case study
The problem
A "Junior Project Coordinator" role asked for 3–5 years' experience, a PMP certification, and advanced Excel — for a job that was really scheduling meetings and updating status reports. A strong candidate read it, assumed they weren't qualified, and never applied. The company never knew why.
The MVP
- Requirement Splitter — forces every requirement into must-have vs. nice-to-have, so teams see when they're over-filtering.
- Bias Detector — flags gender-coded and elitist language, with plain reasons and alternatives.
- Fairness Score — one 0–100 signal across language, requirements, transparency, and readability.
The risk I'd lose sleep over
That FairPost becomes a compliance shield — a high score that lets companies look fair on paper while biased hiring continues behind it. My guardrail: track downstream hiring outcomes, not just posting quality.
What I learned
An AI product's success has less to do with model quality than with whether it changes human behavior. The real challenge isn't intelligence — it's placement in the workflow, incentives, and enforcement.
View the full deck (PDF) →