Collaborations
The discipline of measurable AI-nativity is bigger than one lab. We're building it in the open — with researchers, universities, and the people who build agents. Three ways in:
For researchers
We publish datasets and preprints, submit to peer-reviewed venues, and hold ourselves to the same strict scale we apply to others (L0.9 / 5 today — self-assessment). The evaluation harness and the L0–L5 rubric are available to researchers on request.
Who we're looking for: collaborators in ML / agent evaluation, learning sciences, organizational psychology, and measurement science. If you evaluate agents, study how organizations learn new practices, or work on the validity of organizational measures — we want to compare notes.
Open directions and all public artifacts (datasets with DOIs, preprints) live on the Research page.
For universities & training providers
Teaching humans and agents together needs an architecture, not a course catalog. HALA — our Human-AI Learning Architecture — is the framework Agents University runs on, and it is built to be universal: the same architecture serves companies, universities, and training programs.
- Run your programs on a measurable architecture — with L0–L5 placement before and after, so learning impact is defensible beyond completion rates.
- Co-develop the human side of the discipline — competency models for human–agent work, validated levels, honest measurement.
Framework details — white paper on request.
For agent builders & vendors
Your customers are starting to ask the question we measure: can this agent be trusted in production? Two ways to work with us:
Benchmark your agent
K-Forge, our evaluation dataset, is public (Zenodo, CC BY 4.0) — you can run it against your own agent today. The coverage map shows where your product sits in the capability tree of an AI-native organization, and where it connects to everything else.
Reliability certification
We assess agent reliability against the same evidence-classed framework we apply to organizations — measured behavior in production, not a marketing badge. The certification offering is in its early-partner phase: we're taking a small number of vendors first.
Something else in mind?
We partner with AI labs, universities, and organizations building the next generation of human–AI systems. If your idea doesn't fit the boxes above — that might be the most interesting kind.
Collaborations. (partners named once confirmed.)