OpenAI launches Presence for enterprise voice and chat agents

OpenAI has launched OpenAI Presence, an enterprise product for deploying AI voice and chat agents in workflows such as customer support and outbound sales, the company said Wednesday. The service is being offered only to eligible enterprises through managed deployments, rather than as a self-serve software product.

Presence reflects OpenAI’s push to move AI agents from general-purpose tools into production systems for large companies, where data access, permissions, testing and operating controls can determine whether automation works at scale. The company said each deployment begins with an agent focused on a single task, then uses post-launch monitoring to identify capability gaps and suggest changes that allow the agent to improve over time.

Managed deployment model

OpenAI said Presence deployments are led by its Forward Deployed Engineers and selected global systems integrators. Under that model, OpenAI works with a customer to identify a workflow with business value, connect the relevant knowledge sources and enterprise systems, set permissions and policies, test the agent and put it into production.

The arrangement is designed for companies that need AI agents to operate inside existing technology environments, rather than as stand-alone chat interfaces. In practice, that can mean giving an agent access to approved internal information, defining which actions it may take and setting boundaries for escalation or human review.

OpenAI said it and selected systems integrators can continue supporting a deployment as it expands. The company did not disclose pricing, contract terms or the full eligibility criteria for enterprises using Presence.

Early use in financial services and insurance

OpenAI said it already uses Presence to run its English-language phone support channel. The product is also being used by BBVA, SoftBank and Insurance Australia Group Limited to explore customer-service applications, according to the company.

Daniel Ordaz, head of AI transformation at BBVA Mexico, said in OpenAI’s announcement that BBVA is working with OpenAI as a design partner for Presence to help develop voice experiences for financial customer service. He said the bank is focused on faster, more seamless and more personalized interactions across customer touchpoints.

The customer-service focus aligns with broader corporate interest in generative AI for contact centres and client support. A PYMNTS Intelligence report, “AI in the Chat: Generative AI’s Growing Role in Customer Service,” found that customer service had become one of the more promising applications for generative AI, including chatbots, service requests, personalized interactions and other customer-facing uses.

Part of a broader enterprise agent strategy

Presence follows OpenAI’s February announcement of Frontier, a platform that combines enterprise software with support from human engineers to help companies build, deploy and manage AI agents. OpenAI said Frontier is intended to provide agents with shared context, onboarding, hands-on learning and defined permissions and boundaries.

The two offerings point to a similar commercial strategy: selling AI agents to enterprises as supported deployments, with engineering work around workflow design, systems integration and governance. For regulated industries and large service organizations, those surrounding controls can be as material as the underlying model, particularly when agents interact directly with customers or handle sensitive operational tasks.

OpenAI did not provide deployment volumes for Presence or financial details for the product. The company framed the launch as part of its effort to help enterprises put AI agents into live operations, beginning with defined tasks and expanding after testing and performance review.

This story draws on original reporting from PYMNTS.