OpenAI just stopped selling API access. Now it's selling managed AI agents that resolve 75% of calls without humans.

OpenAI launched Presence on July 22.

Not a new model. Not a new API tier.

A fully managed enterprise agent platform that resolves 75% of inbound support calls without a human touching them.

Here's what Presence actually bundles:

Policies. Guardrails. Permission controls. Escalation rules. Evaluation tools. Simulation testing. And a Codex-powered improvement loop that proposes changes to agent behavior based on real production sessions.

Every deployment starts with one job. Resolving billing disputes. Processing insurance claims. Handling IT tickets.

The agent gets only the knowledge and system access needed for that specific task.

OpenAI is already running it on their own English-language phone support. 75% auto-resolve within weeks. Human handoffs dropped 15 percentage points in 10 days after Codex tuning.

Design partners include BBVA for AI voice support in Mexico, SoftBank for Japanese-language customer conversations, and Australian insurer IAG for surge support during severe weather events.

This is not self-serve. No signup page. No API key.

Deployments are led by OpenAI Forward Deployed Engineers and a small set of global systems integrators.

Your vendor strategy just narrowed. If you were building custom agent orchestration layers on top of OpenAI's API, OpenAI is now competing with you directly.

Audit your agent architecture today. The companies building middleware on raw model access just lost their moat.

SOURCE: https://www.helpnetsecurity.com/2026/07/22/openai-presence-ai-agent-platform/
VERIFIED: Help Net Security, The Robotics Media, OpenAI launch announcement
SIGNAL: OpenAI's shift from model provider to managed agent platform vendor reshapes the entire enterprise AI middleware market.