OpenAI is trying to solve the mess companies hit after the demo: Presence is a governed enterprise product for deploying and managing realtime voice and chat agents, with policies, evaluations, guardrails, and escalation rules baked into the operating system of the agent.

The big catch is how it gets deployed. Presence is not self-service. OpenAI says it is available immediately through a limited general availability program, with OpenAI Forward Deployed Engineers (FDEs) and select global systems integrators leading deployments. That matters, because in enterprise AI, the hard part is usually not getting a model to answer questions. The hard part is getting it to follow your rules, use your systems safely, and degrade gracefully when the real world gets weird.

Presence is designed for eligible enterprise customers that want agents to answer questions, access company systems, take approved actions, and escalate to human workers while operating under company-defined policies, permissions, and evaluation standards. In plain terms: the enterprise controls what the agent is allowed to do, what it can do automatically, what requires approval, and when a person has to take over. OpenAI also positions Presence as a way to package the “stitching” enterprises typically have to build themselves: policies, system connections, evaluations, guardrails, and the update processes that keep an agent aligned over time.

OpenAI addresses a common buyer concern up front. What if an enterprise does not want to rely only on OpenAI models? An OpenAI spokesperson clarified that "Presence uses OpenAI models for the core agent, while allowing customers to connect third-party models and services through APIs for guardrails, tools, and other parts of their workflow." OpenAI has not disclosed pricing, geographic limits, contractual terms, or the expected cost of the engineering and integration work that typically accompanies a deployment, and the source notes the author asked OpenAI about pricing twice and is awaiting a response.

Why all this governance theater? Because enterprises are moving beyond “AI demonstrations” toward production workflows where business rules, customer needs, and operating conditions change. Presence is explicitly framed as a response to that reliability problem. Each deployment starts with a defined job, like resolving a billing issue, supporting an insurance claim, or handling an employee IT request. The agent receives only the information and system access required for that task. Before going live, teams can test agents against common requests, unusual edge cases, and higher-risk scenarios. Graders evaluate whether the agent reached the intended outcome, followed policy, used tools correctly, and escalated when required. Guardrails can intervene when interactions move outside the organization’s defined boundaries.

OpenAI also describes monitoring after launch. Production sessions, escalations, and quality signals can show where an agent is working as intended and where it needs attention. OpenAI says Codex, using a Presence plugin, investigates those signals and proposes updates. Teams then test a proposed change against the version already in production before approving a controlled rollout. The point is to address a brutal enterprise reality: an agent that works at launch can become less reliable when policies, products, or user behavior change. Presence is intended to give enterprises a formal mechanism to update behavior without letting an automated system rewrite itself unchecked.

There is already at least one live use case. OpenAI says Presence powers its English-language phone-support channel at 1-888-GPT-0090, handling open-ended requests, verifying callers, using account context, and performing approved actions. OpenAI also claims the system now resolves 75% of inbound issues without human assistance, and that its Codex-powered improvement loop reduced human handoffs by 15 percentage points over a 10-day period. Those figures are company-reported and have not been independently verified by the source.

Presence is also gaining interest from large organizations. BBVA is exploring voice support for routine banking needs in Mexico. SoftBank is testing natural Japanese-language customer conversations, and Australian insurer IAG is exploring support during high-demand periods like severe weather and natural disasters. The source includes statements from Daniel Ordaz, head of AI transformation at BBVA Mexico, and Tadahisa Murakami, vice president and head of the Data & Digital Transformation Division at SoftBank Corp, each describing collaboration with OpenAI and the goal of “trusted customer agents” that can naturally communicate, connect to processes, and represent the company consistently across interactions.

Finally, Presence signals something broader about where enterprise AI budgets are going. OpenAI is expanding its enterprise strategy beyond APIs and subscription software by formalizing a high-touch deployment model. Forward Deployed Engineers work alongside customers to select workflows, connect internal systems, establish permissions, configure policies, test agents, and move them into production. The delivery approach resembles Palantir’s forward-deployed model in terms of embedding technical personnel close to the customer’s operations, though the source notes the products are not interchangeable: Palantir has historically centered on data integration, ontologies, and operational decision systems, while Presence is more narrowly focused on AI-agent behavior, approved actions, evaluations, escalation, and continuous improvement.

This week’s launch also sits in a wider competitive shift. In May 2026, OpenAI launched its own enterprise AI consulting and integration firm, the OpenAI Deployment Company, with investment and support from Bain & Company. It also offers programs for model customization and fine-tuning. Meanwhile, Anthropic has moved toward a services-led enterprise model through Ode, its consulting organization built around forward-deployed engineers helping companies integrate Claude into complex workflows, which launched just a week prior. The common thread for decision-makers is straightforward: enterprises often need more than access to a model. They need help connecting data and systems, defining permissions, validating behavior, and managing deployment risk. Presence is OpenAI’s explicit answer, and other buyers will have to decide whether they want to keep building governance from scratch or buy it as a product backed by engineers.