- OpenAI launched Presence on July 22 as a managed enterprise platform for deploying production voice and chat AI agents, with no self-serve option available [1] - BBVA Mexico, SoftBank Corp., and Australian insurer IAG are among the first enterprise design partners using the platform [2] - OpenAI's own phone support line, powered by Presence, resolves 75% of inbound issues without human intervention, and a Codex-powered improvement loop cut handoffs by 15 percentage points in 10 days [3] - Presence sits within OpenAI's broader enterprise push anchored by DeployCo, a $4 billion deployment subsidiary launched in May 2026 at a $10 billion pre-money valuation
[4] OpenAI on July 22 launched Presence, a managed enterprise platform that lets large organizations deploy AI-powered voice and chat agents into production customer service and internal workflows. The product is the company's most direct move yet into the business of running AI agents for paying customers, not just selling model access [1].
Unlike OpenAI's existing ChatGPT Workspace Agents — a self-serve builder for internal use — Presence targets external, customer-facing deployments and requires hands-on implementation by OpenAI's Forward Deployed Engineers or select systems integrators. BBVA Mexico, SoftBank Corp., and Australian insurer IAG are among the first design partners [2].
Pricing has not been disclosed. OpenAI described Presence as available in "limited general availability" for qualifying enterprise customers, with broader access and pricing to follow [3].
What Presence Does #
Presence is a deployment and management layer that sits on top of OpenAI's models — including GPT-Live and GPT-5.6 for real-time voice — and handles the operational complexity of running AI agents in production. The platform includes several governance components: policy and SOP definition to control agent behavior, guardrails that automatically intervene when conversations exceed company-defined boundaries, scoped tool access to enterprise systems, and simulation and evaluation tooling for pre-launch testing [2].
A key differentiator is the Codex-powered improvement loop. After launch, the system reviews production sessions and proposes behavioral updates, which are tested before deployment. OpenAI said this mechanism reduced human handoffs by 15 percentage points within 10 days on its own support line [3].
Agents built on Presence can verify identity, access account data, apply policies, and take approved actions such as issuing refunds — then escalate to human employees when cases exceed their scope [2].
Early Customers and OpenAI's Own Deployment #
OpenAI is using Presence to power its own English-language phone support line at 1-888-GPT-0090, where it reports that approximately 75% of inbound issues are resolved without human intervention [3].
Among external partners, BBVA Mexico is exploring AI voice agents for banking support, SoftBank Corp. is testing Japanese-language agents for frontline customer teams, and IAG — one of Australia's largest insurers — is deploying support agents for high-demand periods such as severe weather events [2].
Notably, both SoftBank and BBVA are also financial backers of DeployCo, OpenAI's $4 billion deployment subsidiary formed in May 2026, suggesting deep strategic alignment between the investment and product relationships [4].
The DeployCo Connection #
Presence is the first major product to emerge from OpenAI's broader enterprise services strategy. In May 2026, OpenAI launched DeployCo, a majority-owned subsidiary backed by $4 billion in investment at a $10 billion pre-money valuation. The unit was funded by a consortium of 19 partners led by TPG, with Advent, Bain Capital, and Brookfield as co-leads, alongside consulting firms McKinsey and Capgemini [4].
DeployCo's model stations OpenAI's Forward Deployed Engineers — bolstered by roughly 150 specialists acquired from consulting firm Tomoro — inside large enterprises to convert corporate data and workflows into production AI systems built exclusively on OpenAI's models [4].
Presence operationalizes that model for a specific, high-value use case: customer service automation. The consulting-style deployment, with no self-serve tier, mirrors the approach enterprise software companies like Palantir have used to embed deeply within client organizations [1].
Competitive Landscape #
OpenAI's entry into managed enterprise agent deployment puts it in direct competition with a growing field. Salesforce has been building its Agentforce platform, Google Cloud offers its own agent-building tools through Vertex AI, and a wave of startups — including Sierra AI, founded by former Salesforce CEO Bret Taylor — are targeting the same customer service automation market.
The managed, white-glove approach differentiates Presence from API-first competitors but also limits its initial scale. Without self-serve access or published pricing, adoption will depend on OpenAI's ability to staff deployments and demonstrate ROI at each customer [1].
Questions remain about compliance. Multiple reports noted that OpenAI has not clarified how Presence handles specific regulatory requirements such as the EU AI Act, a gap that enterprise buyers in regulated industries will need addressed before committing to production deployments [1].
What's Next #
OpenAI has indicated that broader pricing and availability details are forthcoming, though no timeline has been set. The company's immediate focus appears to be on demonstrating production results with its initial cohort of design partners while scaling its Forward Deployed Engineering team [2].
For a company that has historically generated the bulk of its revenue from consumer subscriptions and API access, Presence represents a significant strategic bet on high-touch enterprise services — a business model with higher margins per customer but fundamentally different unit economics than selling API calls [4].
Companies mentioned #
Further sources #
[1] The Decoder — OpenAI Presence wants to make AI agents production-ready for busi… ↗ [2] The SaaS Sentinel — OpenAI Launches Presence, an Enterprise Platform for Voice … ↗
[[3] ExplainX — OpenAI Presence Enterprise Voice Chat Agents July 2026 ↗](https://explainx.ai/blog/openai-presence-enterprise-voice-chat-agents-july-2026)
[[4] Axios — OpenAI launches AI consulting arm valued at $14 billion ↗](https://www.axios.com/2026/05/11/openai-deployco-private-equity)
The stories that matter, in one email. Free — unsubscribe anytime.