Overview
- OpenAI announced Presence on Wednesday, July 22, 2026, describing it as a deployment platform that connects AI agents to a company’s internal data, policies, software, and workflows.
- Presence requires case-by-case engagements with OpenAI engineers or designated partners rather than self-serve access, with limited general availability supported by forward-deployed engineers and select partners.
- The platform includes prelaunch simulations, automated graders, guardrails, human review and policy controls, and uses OpenAI’s Codex to analyze live interactions and suggest approved updates after launch.
- OpenAI says it already uses Presence for its English-language phone support and that the system resolves about 75% of inbound calls without human intervention.
- The launch signals a strategic shift from selling raw models toward higher-margin, integrated enterprise software but also raises competition and governance concerns because some customers build similar agent products on OpenAI models.