OpenAI Presence is a new AI agent platform built for customer support

OpenAI just launched Presence, an enterprise platform that helps companies put AI voice and chat agents to work on real customer conversations. It answers questions, resolves issues, and hands off to a human when the situation calls for it. OpenAI announced the product on July 22, 2026, and it already powers the company’s own English-language phone support line. If you run a business and wonder whether AI customer support is ready for prime time, OpenAI Presence is the clearest signal yet that the answer is yes, at least for the companies with the budget to deploy it properly.

What OpenAI Presence actually does

Presence pairs a model with company policies, guardrails, and escalation rules. Each deployment starts with one job, like resolving a billing issue or answering an insurance claim question. The agent gets only the data and system access it needs for that job, nothing more. The business decides what the agent can do on its own and when it needs to bring in a person.

Before any of this reaches a real customer, teams run simulations against common requests, tricky edge cases, and high-risk scenarios. After launch, OpenAI’s Codex tool watches production sessions and escalations, then proposes fixes that a human reviews and approves. This closes a loop that most chatbot projects never had: someone is always watching for the moment the agent gets confused, and something can actually be done about it.

The results so far back up the pitch. Presence runs OpenAI’s own support line and now resolves 75 percent of inbound issues without a human. In one ten-day stretch, the Codex improvement loop cut human handoffs by 15 percentage points. Early customers include BBVA Mexico, which is testing voice support for everyday banking, and SoftBank, which reports strong marks for natural Japanese-language conversations. You can read the full rundown in OpenAI’s official Presence announcement.

Why this matters even if your business is nowhere near enterprise scale

Presence is not available as a self-serve product yet. OpenAI works directly with each customer through its Forward Deployed Engineers, and that kind of hands-on rollout points at large budgets, not a five-person shop. So why should a small business owner care?

Because Presence shows where the ceiling is, and that ceiling keeps rising fast. A year ago, AI support agents mostly handled simple FAQs. Now one is running a real support line and resolving three out of four tickets on its own. VentureBeat’s coverage of the launch notes that OpenAI built Presence on lessons from years of enterprise deployments, which means the techniques behind it, tight scoping, simulation before launch, and constant review after, will trickle down into cheaper tools within a year or two. This is the same pattern you saw with GPT-5.6’s release: features that debut at the enterprise tier usually show up in accessible form soon after.

What you can do with today’s tools

You don’t need Presence to start improving how AI handles your customer conversations. Give any chatbot a narrow job instead of a vague one. Write down what it’s allowed to decide and what it must hand off to you. Test it against your worst-case customer questions before it goes live, not after. These are the exact principles Presence runs on, just without the enterprise price tag. If you’re weighing which AI provider to build on, our breakdown of Claude Fable 5’s pricing changes is a useful comparison point, and our look at what the 2026 data shows on AI and small business growth covers where owners are actually seeing returns right now.

Frequently Asked Questions

What is OpenAI Presence?

OpenAI Presence is an enterprise platform, launched July 22, 2026, that helps companies deploy AI voice and chat agents for customer support, sales, and internal workflows. It combines OpenAI’s models with company-specific policies, guardrails, and escalation rules so the agent behaves predictably in production.

Can small businesses use OpenAI Presence?

Not yet. Presence is available only to eligible enterprise customers through a limited program, and deployments are led by OpenAI’s own engineers and select partners. It is not a self-serve product, so smaller businesses will need to rely on existing chatbot and voice AI tools for now.

How does OpenAI Presence handle mistakes?

Presence tests agents through simulations before launch, then monitors real conversations and escalations afterward. OpenAI’s Codex tool flags patterns that suggest the agent needs adjustment, and a human team reviews and approves any change before it goes live.

What OpenAI Presence means for the future of business AI

OpenAI Presence is not a tool you’ll install this quarter unless you run a large operation with an OpenAI account team on speed dial. But it’s a preview of where customer support is headed, and it’s moving faster than most owners expect. Watch how the mid-market and small business tools built on GPT and Claude models evolve over the next year. The gap between what enterprises get first and what everyone else gets eventually keeps shrinking, and OpenAI Presence just moved that line again.