Perfolio.AI.
By Aboli, Dan, Rahul & Yujin
Perfolio.AI is an evidence-based performance review tool that automatically extracts impact from work artifacts — PRDs, design docs, Jira tickets, GitHub activity, monthly reports — and maps each output to the company's competency framework, eliminating the seasonal scramble that defines self-review season for knowledge workers.
Built B2B2C (sold to employers, used by employees), Perfolio targets the Busy Knowledge Worker IC who wants their work communicated clearly and credibly when promotion calibrations happen. Instead of writing self-reviews from memory the night before submission, employees upload artifacts → the AI extracts evidence with competency labels as constrained output categories → the user confirms each mapping before narrative generation begins.
The workflow runs in three "trust-by-design" stages: Extract + Map (structured JSON extraction with competency labels), Generate (narrative output from confirmed evidence, with tone and citation), and Confirm (human-in-the-loop validation). Powered by OpenAI GPT-4.1 (with GPT-4.1 mini for lighter tasks) for reliable structured JSON output, the system hit a 78% pass rate on automated evaluations for extraction + competency mapping.
Positioned for the fast-growing AI HR Tech segment — overall HR Tech at 8–12% CAGR, AI in HR at 25–35% CAGR, generative AI productivity tools at 30%+ — Perfolio.AI bets that the future of performance reviews is artifact-grounded, not memory-grounded.