Updates from micro1
Data lab to train frontier models and evaluate agents through expert human data and real-world environments.
- micro1 x DataSnipper: Testing AI Agents without Client Data (Partnership · October 31, 2026)
The company partnered with DataSnipper to create a synthetic evaluation pipeline for auditing agents, allowing for rigorous testing without using sensitive client files. - PII Transformation for Enterprise Datasets (Product launch · September 24, 2026)
The company introduced flow-transform 1.0, a framework designed to convert real identities into synthetic ones while maintaining the semantic utility and consistency of enterprise datasets. - AI Models Now Introduce Safety Risks in the Physical World (Project update · September 18, 2026)
The company's robotics division published research on the safety risks of giving frontier AI models direct control over physical hardware and machines. - $1T/year evals market emerging (Event · January 23, 2026)
The company hosted a virtual discussion featuring the CEO of Box to explore how enterprise AI agents transition from experimentation to production.