OpenAI Presence brings managed voice and chat agents to enterprise workflows

From agent demos to production workflows

OpenAI introduced Presence as a managed enterprise product for deploying AI agents into customer and internal workflows. Rather than exposing a self-serve agent builder, Presence combines model reasoning with deployment support, policies, guardrails, approved actions, evaluations and escalation rules. Each deployment begins with a defined job, and the agent receives only the knowledge and system access needed for that workflow.

Controls around actions and escalation

Companies define what an agent may do, when approval is required and when a person must take over. OpenAI says teams can test common requests, edge cases and higher-risk scenarios before launch, then use production sessions and escalation signals to identify gaps. Codex can propose changes, but teams can test and approve those changes before controlled rollout.

Voice and chat availability

Presence supports real-time voice and chat experiences such as customer service, outbound sales and higher-risk internal workflows. OpenAI says deployments are led by its Forward Deployed Engineers and select global systems integrators. The product is available to eligible enterprise customers through a limited general-availability program and is not currently self-serve.

Why it matters

The announcement reflects a broader shift from standalone AI assistants toward governed agents that can interact with business systems and take approved actions. The important operational distinction is that access, policies, evaluations and human escalation are treated as part of the deployment system rather than relying on model prompting alone.