OpenAI Presence Sells Enterprise AI Agents with Engineers Attached
OpenAI launches a managed enterprise AI product sold as custom projects backed by its own engineers and global integrators.
Stock photo for illustration only, not from the actual event
- OpenAI introduces Presence, a managed enterprise AI agent service.
- Sold as custom projects rather than self-serve software.
- Features a six-stage implementation process from scoping to post-launch.
- Initial named enterprise customers include BBVA, SoftBank, and IAG.
The newest way to buy enterprise AI agents from OpenAI does not involve buying anything online. Announced on July 22, OpenAI Presence is a managed product delivered through a limited general availability programme. The company states plainly that it is not yet available as a self-serve product, with deployments led directly by OpenAI’s own Forward Deployed Engineers and selected global systems integrators.
This marks a notable departure for a business that has historically relied on API keys and seat licences. Presence is sold as a project rather than a standalone product. Each engagement kicks off with a specific job, such as resolving a billing dispute, managing an insurance claim, or clearing an employee IT service request.
Agents are granted only the necessary knowledge and system access required for the specific task. Customers write the governance rules detailing what the agent can do, when sign-off is required, and when a human must take over. Following launch, Codex reads production sessions and escalations to propose changes that the customer’s team tests and approves before rollout.
By deploying its own engineers directly into client operations, OpenAI is addressing a hard truth discovered by enterprises over the past two years: the primary hurdle in production AI is not model capability, but rather integration, permissions, and operational change management.
OpenAI’s documentation outlines a rigorous six-stage delivery process covering human involvement, which includes:
- Scoping business outcomes
- Security, privacy, and legal review
- Simulation and acceptance testing
- Staged rollout
- Post-launch iteration
These measures directly tackle Gartner’s warning that over 40% of agentic AI projects will fail by the end of 2027 due to governance issues and undefined business value. Pre-launch simulations evaluate whether an agent reaches correct outcomes, adheres to policy, and escalates appropriately before interacting with external users.
Stock photo for illustration only, not from the actual event
“OpenAI Presence is battle-tested, assembled from years of deploying agents with enterprise customers before it was packaged and named.”
OpenAI
The initial customer cohort highlights early-stage deployments: BBVA is exploring voice support for everyday banking in Mexico, SoftBank is testing Japanese-language conversations, and IAG is exploring support during high-demand events like severe weather. Daniel Ordaz, head of AI transformation at BBVA Mexico, describes the bank as a design partner helping shape voice experiences for financial customer service.
Pricing remains unpublished, with implementation scope and costs set per customer and deployment. The underlying models are not explicitly named; documentation notes that Presence uses OpenAI models with configurations tailored to specific workflows and subject to evolution as those workflows mature.
Source: AI News
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment