OpenAI launches Presence for enterprise agent deployments
The new service is sold through OpenAI’s deployment teams, not as a self-serve API, with pricing scoped customer by customer.
By Dominic Okoye · Staff Writer
· 3 min read
OpenAI has introduced Presence, a web service for enterprises deploying AI agents across voice and chat, with pricing undisclosed and implementation handled through its deployment teams rather than a self-serve product. The launch pushes OpenAI further into consulting-style enterprise work, where agent rollouts often require integration, governance controls and human fallback rather than API access alone.
OpenAI said Presence is aimed at real-time business workflows including customer support, outbound sales and internal processes with higher risk. A company could use the service to handle a billing request, verify a customer, retrieve account details, apply company policy and take an approved action, according to OpenAI.
The product includes controls for how agents behave, including policy and governance settings. OpenAI said the interface includes an agent editor, a testing playground and simulation tools for scenarios such as refunds, order status checks, account deletion and other administrative tasks.
Presence is being distributed through OpenAI Deployment Company and Forward Deployed Engineers, along with selected global systems integrators. OpenAI said the service is “not yet available as a self-serve product.” The deployment group was introduced last year and expanded after OpenAI acquired consultancy Tomoro in May.
The go-to-market choice is the relevant part for enterprise buyers. OpenAI is not presenting Presence as another API endpoint for internal teams to wire up on their own. It is packaging model capability with implementation labor, closer to the Palantir-style field engineering motion that large companies often require for software touching customer accounts, payments or regulated workflows.
OpenAI said it already uses Presence for its AI English phone support channel. The company said the system can interpret open-ended requests, verify identity when needed, use account context and take approved actions, including billing resolutions, account changes and eligible refunds. OpenAI also said the agent can bring in a human when a request requires it.
That handoff will be a key test. Gartner said last month that by 2027, half of organizations planning to move customer service to AI will abandon those plans. Kathy Ross, a senior director analyst in Gartner’s customer service and support practice, said AI has potential in service operations but “is not a panacea,” adding that many interactions still require human empathy and understanding.
OpenAI also pointed to SoftBank as a customer exploring the service. SoftBank has committed $60 billion to OpenAI, and Tadahisa Murakami, vice president and head of SoftBank’s data and digital transformation division, said the companies are testing how Presence can support customer agents that speak naturally and connect to business processes. He said SoftBank frontline teams rated the agent’s Japanese-language conversations highly for naturalness and accuracy.
OpenAI did not publish list pricing. A spokesperson said that during the limited general availability phase, deployments are scoped individually based on the customer’s use case and implementation needs, with broader pricing details to come as availability expands. OpenAI said Presence is available to eligible enterprise customers and that it will continue supporting existing voice customers that use its models through APIs.
This story draws on original reporting from The Register.
