OpenAI shipped Presence on July 22, its biggest enterprise agent play of the year. It’s a platform for building, deploying, and managing realtime voice and chat agents — the kind that pick up your support call, look up your account, and actually do something about it instead of reading a script.
What it actually is
Not a chatbot you drop on a website. Presence bundles the whole production stack: policies and SOPs, guardrails, approved actions, simulations, evals, and a Codex-powered loop that keeps tuning the agent after launch. It connects to your real company systems, so the agent can take approved actions and escalate to a human when it hits a wall. Target jobs: support, sales development, procurement, IT, HR.
Why it’s worth watching
The proof is internal — Presence runs OpenAI’s English phone line and resolves 75% of calls with zero human touch, up 15 points in 10 days thanks to that Codex feedback loop. Early customers include BBVA Mexico, SoftBank, and IAG.
The catch: it’s not self-serve. OpenAI Forward Deployed Engineers and select integrators lead every deployment, so there’s no public API to sign up for — you get a team, not an endpoint. That’s OpenAI going straight after the enterprise CX market, Palantir-style.
You Might Also Like
- Openai Realtime Voice Webrtc Stack the Infra Blueprint Every Voice Agent Startup now has to Compete With
- Openai the Development Company 4b Raise to Embed Engineers Inside 2000 mid Market Firms
- Openai Acquires ona Formerly Gitpod to Keep Codex Agents Coding While Your Laptop Sleeps
- Openai gpt Realtime 2 1 gpt Realtime 2 1 Mini cut Voice Agent Latency by 25
- Happycapy Might be the Easiest way to Actually use ai Agents Right now
Top AI Product 
Leave a comment