Coworker launches OM2: enterprise memory layer + model routing claims 51x cost cut

FellMentKE · x · 2026-09-03

Coworker AI launched OM2, a permanent memory layer for enterprise AI agents. Today's agents re-search Slack, docs, and CRM from scratch every task — over 50% of the token bill goes to searching for context, not the answer. OM2 gives AI a persistent memory of the company; the company claims it makes enterprise AI 9x cheaper, and 51x with model routing that sends each request to the right model, with 84.5% better quality (per their benchmarks vs. Claude native tooling). Available via Coworker MCP (works in Claude Code, Cursor, any MCP client, 50+ tools) or native apps; SOC 2, 50+ connectors.

Original post →

More from coding & agent

coding & agent channel →