Anthropic and Accenture create an embedded-evaluator track inside Anthropic

Anthropic and Accenture announced a partnership to build embedded evaluators who will work inside Anthropic with employee-like access while red-teaming models and testing safeguards.

Editorial illustration of human evaluators reviewing AI model outputs in a lab setting.

Anthropic and Accenture announced a partnership on September 18, 2026, to build a team of embedded evaluators who will work inside Anthropic. Anthropic said the evaluators will have access comparable to an employee’s while they red-team models, conduct alignment assessments, and test safeguards.

Accenture said each company expects to invest at least $1 billion over five years. Anthropic also said the partnership is non-exclusive and that more evaluators will be announced in the coming weeks.

The supplied sources do not disclose headcount, job titles, or any broader staffing plan beyond the embedded-evaluator track.

The bundle also includes an Associated Press report that Anthropic said Claude is leading 26% of its model research and development, and that about 90% of Anthropic’s research and development is done in collaboration with Claude. The AP also reported that the company said it is setting up external third-party evaluators embedded within the company.


Sources