OpenAI unveils Presence, a new platform that lets enterprises launch and manage realtime voice agents and chatbots
VentureBeat ● Covered by 4 sources
OpenAI launched Presence, a platform for enterprises to build and manage voice agents and chatbots at scale. It's OpenAI moving further into the corporate software stack, not just chat.
Based on reporting by VentureBeat — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI wants a bigger seat at the enterprise table, and Presence is the pitch. The new platform lets companies stand up voice agents and chatbots for both customer-facing work and internal operations, all wrapped in whatever rules, permissions, and evaluation standards a business wants to enforce. Think call centers, internal help desks, onboarding flows — anywhere a company might have once hired a human to answer the same fifty questions on repeat.
The product is currently limited to eligible enterprise customers, which tells you OpenAI is being deliberate about the rollout rather than throwing it open to anyone with an API key. That caution makes sense. Voice agents that interact with real customers or handle internal business logic carry more risk than a chatbot answering trivia in a sidebar. Get the guardrails wrong and you're looking at a very public, very embarrassing failure — the kind that ends up as a screenshot on social media within the hour.
What's notable here is the emphasis on company-defined policy and evaluation, not just raw model capability. OpenAI has spent the last couple of years selling the idea that its models are smart enough for anything. Presence is a quieter admission that smart isn't the same as deployable. Enterprises don't just want a capable model; they want control over what it's allowed to say, who it can talk to, and some way to check its work after the fact.
This also puts OpenAI in more direct competition with the growing pile of startups building voice-agent infrastructure — companies that have spent the last two years building exactly this kind of permissioning and evaluation tooling as their entire business. If OpenAI folds that functionality into its own platform, it's a rough morning for anyone whose pitch deck says 'the layer between you and OpenAI's models.'
My take — AI-written commentary, not fact-checked reporting
This is OpenAI doing what every platform company eventually does: eating the middleware. If your startup's whole value proposition was 'we make OpenAI's models safe and controllable for enterprises,' you should be nervous, because that's now a checkbox in Presence's settings menu. I don't think this is a step toward AGI or anything — it's a very normal, very smart platform play, and it'll probably work.
Read more about this at: VentureBeat