# PYX Labs

> PYX Labs is a research lab sponsored by Perceptyx focused on defining evaluation standards for how AI systems understand, interpret and reason about people in the workplace. The lab leverages proprietary employee experience datasets, expertise in behavioral science and organizational psychology, and structured evaluation methodologies to assess AI performance in real-world organizational contexts. PYX-Voice is the first in a planned series of benchmarks designed to help the AI industry build systems that are not only capable, but trustworthy in their understanding of employees.

Key facts:
- Flagship benchmark: PYX-Voice, which measures how well frontier AI models perform real employee-listening analysis (interpreting employee survey and feedback data), scored against expert criteria across 84 validated tasks with a public leaderboard.
- Methodology: five steps from real workplace data to a defensible score; criteria authored by practicing I-O psychologists; graded by a cross-vendor judge; results validated against an expert bar.
- Audience: frontier AI labs, workplace/HR AI product builders, academic researchers, and enterprises evaluating AI for employee-facing work.
- Sponsor: Perceptyx (perceptyx.com). Academic advisors at Stanford University and UT Austin.

## Benchmarks

- [Benchmarks overview](https://www.pyxlabs.ai/benchmarks): All PYX Labs benchmark standards for AI work that touches employees
- [PYX-Voice](https://www.pyxlabs.ai/benchmarks/pyx-voice): Live leaderboard — how frontier models (GPT-5, Gemini, Claude, Grok) score on real employee-listening analysis, 84 tasks

## Research

- [Research](https://www.pyxlabs.ai/research): The PYX-Voice research paper (v1.0) — how the benchmark was built, how frontier models scored, and what they consistently miss
- [Methodology](https://www.pyxlabs.ai/methodology): The five-step standard behind every PYX Labs benchmark, from real data to a defensible score

## Company

- [About PYX Labs](https://www.pyxlabs.ai/about): What PYX Labs does — benchmarks, expert post-training data, and evaluation — and who is behind it
- [Contact](https://www.pyxlabs.ai/contact): Get in touch — AI labs, HR-tech vendors, researchers, and press

## Optional

- [Home](https://www.pyxlabs.ai/): PYX Labs homepage