OpenAI has launched Presence, a new enterprise platform designed to help organizations deploy battle-tested AI agents capable of answering questions, resolving issues, accessing company systems, taking approved actions, and escalating to human staff when needed. The platform, announced July 22, 2026, represents OpenAI’s most ambitious entry into production AI agent deployment.
The platform addresses a critical challenge facing enterprises: moving AI agents from proof-of-concept to reliable production deployment. According to OpenAI, the focus has shifted from proving agents can work to making them reliable enough for high-value work in real-world environments.
“Agent behavior must adapt as products, policies, and user behavior change,” OpenAI explained in its announcement. “That requires more than a model: it requires the systems, evaluations, and deployment expertise to improve agents as those conditions change without giving up control.”
Each Presence deployment starts with a specific job — such as resolving billing issues, supporting insurance claims, or handling IT service requests. The agent receives only the knowledge and system access required for that particular task. Companies set the policies defining what the agent can do, when it needs approval, and when a human should take over.
The platform has already proven itself in demanding environments. OpenAI’s own English-language phone support channel (1-888-GPT-0090) runs on Presence, handling open-ended requests, verifying callers, using account context, and taking approved actions. Within weeks, it met or exceeded benchmarks used to grade frontline human support quality and now resolves 75% of inbound issues without human assistance.
A Codex-powered improvement loop reduced human handoffs by 15 percentage points in just 10 days, demonstrating the platform’s ability to continuously refine agent performance based on production feedback.
Several major enterprises are already building on Presence:
- BBVA is exploring AI-powered voice support for everyday banking needs in Mexico
- SoftBank is testing natural Japanese-language customer conversations
- IAG is exploring timely support during high-demand events such as severe weather
The platform brings together the components enterprises need for production agent deployment: policies and standard operating procedures, guardrails, approved actions, simulations, evaluation tools, and a Codex-powered improvement process. Teams can test deployments against common requests, edge cases, and higher-risk scenarios before going live.
Post-launch, production sessions, escalations, and quality signals reveal where agents work well and where they need attention. The Codex-powered improvement system investigates those signals and suggests updates, which teams can test against the production version before approving controlled rollouts.
Presence is available to eligible enterprise customers through a limited general availability program, with deployments led by OpenAI Forward Deployed Engineers and select global systems integrators. The platform currently supports real-time experiences across voice and chat channels, including customer support, outbound sales, and high-risk internal workflows.