OpenAI Presence puts enterprise agents behind a managed gate

Jul 22: OpenAI introduced Presence, a managed product for deploying voice and chat agents in enterprise workflows. It bundles company context, permissions, approved actions, simulations, graders, guardrails, human escalation, and a Codex-powered improvement loop. 1

This is an operational release, not a new model. Presence sits around OpenAI models including GPT-5.6 and GPT-Live, turning the API-first baseline into an OpenAI-led implementation with production controls. VentureBeat reports that customers can connect third-party models and services through APIs for tools and guardrails, while the core agent uses OpenAI models. 2

OpenAI says its English-language phone line, 1-888-GPT-0090, resolves 75% of inbound issues without human assistance. Its Codex loop cut human handoffs by 15 percentage points in 10 days. Three design partners, BBVA Mexico, SoftBank Corp., and IAG, are exploring banking, Japanese-language support, and insurance workflows. These are vendor-reported outcomes or partner evaluations, not an independent benchmark. 1 No Jitter repeats the figures without independently validating them. 3

Presence is available to eligible enterprise customers through limited general availability. OpenAI Forward Deployed Engineers and selected systems integrators lead rollouts; it is not self-serve. Pricing, geographic limits, contract terms, and implementation cost remain undisclosed. No Jitter says deployments are scoped individually. 23

Use the existing API if you need a testable path today; contact sales if you want OpenAI to own more of the deployment work. Would you pilot a 75%-resolution agent behind this managed control plane, or build the workflow around GPT-5.6 and your own guardrails?