A research lab, not a product demo

We study how organizations think — and build the substrate for it to think correctly.

Reflective Labs is a collective intelligence and organizational learning lab. We publish primary research — with proofs, benchmarks, and named limitations — on the question of how a group of humans and machine agents can reach decisions that are deterministic, auditable, and provably bounded, instead of merely plausible.

9 papers published · updated monthly · full PDFs and source on request

What we are

A collective intelligence, organizational learning lab.

Most AI research asks how a single model can get smarter. We ask a different question: how does a group — of people, of agents, of solvers with different kinds of certainty — reach a decision that the group can actually stand behind? That is collective intelligence in the literal sense: not one model reasoning alone, but many differently-capable reasoners converging on a shared, governed answer.

The organizational learning half is the harder discipline. A decision is not just made once — it is revisited, contested, and its outcome feeds back into how the next decision gets made. We study the loop: how authority narrows as intent decomposes, how disagreement gets institutionalized rather than smoothed over, and how a system can learn better priors without ever letting learning quietly become authority.

It was never only about language models. Our engine is a mixture of experts in the 1991 sense — heterogeneous solvers behind a policy gate, not a learned softmax inside one network. A language model is an excellent participant. It is one participant among several, alongside constraint solvers, SMT search, closed-form statistics, and fitted models — each admitted through the same gate, on the same terms.

The archive

The research library.

Every paper is a real technical report: propositions, proofs, benchmark numbers, and a named limitations section. Each has a downloadable PDF and a short-form slide summary. Source behind every claim is shared on request.

Browse all 9 papers →

The applications built on this research live at Reflective Group.

Axioms™, Helms™, Converge™, and Organism™ — the product layer that puts this research to work in real organizations.

Visit the company site →