Explore / AI Agent Research
AI Agent Research
A clear evidence trail matters more than a synthetic agent score.
Published collection
Start here
Every card opens a complete page with clear evidence and limitations.
01research
How AI Dots Tests and Verifies AI Agents
Every capability claim needs a source and a state. A vendor claim and an independent observation are different kinds of evidence.
Explore02comparison
OpenAI Dots vs Meta Muse
Both vendors describe personal agents that can continue work between conversations. Their documented setup, access, and controls differ.
ExploreCurrent evidence
The launch profiles cite official documentation. No hands-on benchmark results have been published by AI Dots yet.
How we test
The methodology defines repeatable tasks, run counts, evidence capture, and limitations before a result can be called Verified by AI Dots.