OpenAI Launches Presence, an Enterprise Platform for Voice and Chat AI Agents

Weeks after expanding testing of its GPT-5.6 models, OpenAI introduced Presence on July 22, 2026, a new enterprise platform designed to help companies deploy voice and chat AI agents across customer-facing and internal workflows. Presence marks a shift in OpenAI's enterprise strategy, moving from offering raw model access toward a fully managed system for building, governing, and continuously improving production-grade AI agents.

OpenAI says its own English-language phone support line already runs on Presence, resolving 75% of inbound issues without human intervention. Early enterprise partners include BBVA in Mexico, SoftBank in Japan, and the Australian insurer IAG, each testing the platform for different customer-service scenarios.

What Happened: OpenAI's Presence Launch

Presence is built around real-time voice and chat interactions that enterprises can deploy across multiple channels, including phone support, in-app chat, and internal IT service requests. Rather than handing a company raw access to a language model and leaving the deployment work to them, OpenAI packages Presence with standard operating procedures, approved action definitions, and simulation and evaluation tools that let businesses test an agent's behavior before it goes live.

Each Presence deployment is scoped to a specific task. Enterprises set policies that define what the agent is allowed to do, when it needs approval before acting, and when a human should take over the conversation entirely. OpenAI also built an improvement loop into the product, using a tool called Codex to propose updates to an agent's behavior based on real production data and any escalations that occurred.

Key Details

OpenAI's own customer support line, which runs on Presence, has resolved 75% of inbound issues without human assistance. The company also reported that its internal improvement cycle reduced human handoffs by 15 percentage points within just 10 days on that same support channel, an indication of how quickly the Codex-driven refinement process can adjust agent behavior based on real usage.

Like Salesforce's own prospecting agents for Sales Cloud, Presence is not being offered as a self-serve product. It is available only to eligible enterprise customers through a limited general availability program, with deployments supported directly by OpenAI's Forward Deployed Engineers and selected global systems integrators. That white-glove approach has led some coverage to describe the launch as OpenAI moving into consulting-style, boots-on-the-ground pricing for enterprise AI agent deployment.

Why It Matters

Presence represents a meaningful shift in how OpenAI positions itself for enterprise customers. Instead of competing purely on model capability through APIs, OpenAI is now competing directly with contact-center and customer-experience software vendors by offering a complete, managed agent deployment system, guardrails, testing tools, and human-escalation policies included.

That positioning also explains why the launch drew attention from investors watching software-as-a-service stocks. Coverage of the announcement noted that SaaS shares tied to customer service and contact-center software dropped following the news, reflecting concern that OpenAI's move into managed enterprise agents could compete directly with established players in that space.

Industry Context

Presence arrives as enterprise adoption of AI agents accelerates across customer service, sales, and internal operations. Early partners give a sense of the range of use cases OpenAI is targeting: BBVA is exploring voice support for everyday banking services in Mexico, SoftBank is testing Japanese-language customer conversations, and IAG is examining how Presence could support customers during high-pressure situations such as severe weather events.

"As a design partner for Presence, we are working with OpenAI to help shape and refine voice experiences for financial customer service," said Daniel Ordaz, Head of AI Transformation at BBVA Mexico. That kind of design-partner language signals OpenAI is still actively shaping Presence's capabilities alongside its earliest enterprise customers rather than shipping a finished, one-size-fits-all product.

What It Means for Users and the Industry

For consumers, Presence's rollout means a growing share of phone and chat support interactions, at banks, telecoms, and insurers among others, may soon be handled by AI agents built on OpenAI's infrastructure rather than by traditional call-center software or human-only teams. The platform's escalation rules are meant to ensure sensitive or complex requests still reach a human, but the overall trend points toward AI handling a larger first-line share of customer interactions.

For enterprise software vendors, Presence is a direct competitive signal. OpenAI is no longer just a model supplier that other companies build customer-service tools on top of, it is now building and selling the full deployment stack itself, which could squeeze out vendors whose main value proposition was wrapping OpenAI's models in their own agent-management tooling.

What Happens Next

OpenAI has not announced a timeline for moving Presence beyond limited general availability into a broader self-serve offering. With early partners like BBVA, SoftBank, and IAG still in active testing and design-partner phases, the coming months are likely to focus on refining the platform based on real deployment data before OpenAI considers expanding access to a wider set of enterprise customers.

Final Takeaway

Presence signals that OpenAI intends to compete directly in the enterprise AI agent deployment market, not just supply the underlying models. With a 75% automated resolution rate on its own support line and design partnerships spanning banking, telecom, and insurance, OpenAI is positioning Presence as a managed alternative to piecing together agent infrastructure from separate vendors, a bet that could reshape how large enterprises think about deploying AI at the front line of customer service.

How Presence Differs from a Standard AI Chatbot Deployment

Most enterprise AI agent deployments today involve stitching together a language model, a set of guardrails, and monitoring tools from several different vendors, then maintaining that pipeline in-house. Presence collapses that work into a single managed system: OpenAI handles the model, the policy layer, the testing environment, and the ongoing improvement cycle, while the enterprise defines the rules the agent must follow and reviews performance data.

That managed approach carries tradeoffs. Enterprises gain faster deployment and a continuously improving agent without needing to build their own AI operations team, but they also become more dependent on OpenAI's infrastructure and roadmap for a system now handling direct customer interactions. For large regulated industries like banking and insurance, where BBVA and IAG are among the first testers, that tradeoff between speed and control is likely to shape how quickly Presence expands beyond its initial design partners.

FAQs