Coworker.ai Launches OM2 Organizational Memory Layer

Coworker.ai launched OM2 on September 3, 2026. The organizational memory layer stores company context so AI models and agents can reuse it instead of retrieving and rebuilding it for every request. The company says this approach reduces token spending on context by 9x.

OM2 ingests data from more than 50 enterprise platforms and converts messages, records, and meeting notes into linked facts that record who was involved, what was decided, and which project or account it touches. Permissions are attached to each fact and connection. The layer works with Claude, ChatGPT, Gemini, Perplexity, and custom agents, and Coworker.ai says it is available headlessly through its Model Context Protocol server as well as through its native apps.

Coworker.ai says OM2 produces responses that users preferred over standard tool calling methods 84.5% of the time and completes tasks 64% faster. Its model routing option can increase the claimed cost reduction to 51x by directing each task to a selected model. More than 300 companies use Coworker.ai, according to the company, including RapidSOS, whose revenue operations director said the team previously spent hours stitching together Salesforce, Slack, and meeting notes to trust an answer.