Presence in one sentence: OpenAI stops selling models and starts selling outcomes
OpenAI Presence is a managed enterprise AI agents platform that deploys realtime voice agents and chatbots directly into customer support, sales, billing, and internal workflows so they can read policies, access systems, take approved actions, and escalate edge cases under company-defined governance—shifting OpenAI’s business from cheap, self-service models to higher-margin deployment and management expertise. That shift is the real story. Presence is not another API; it is OpenAI’s declaration that the value in enterprise AI is no longer the model, but the work the agent owns. By inserting a managed layer between frontier models and everyday workflows, OpenAI is telling CIOs: stop bolting AI features onto old dashboards and start letting agents run the process. For any team that earns renewal fees by "owning" a workflow screen, that is an existential change, not a minor product update.
From API vendor to deployment partner: why Presence is a consulting play
Presence makes it easier for enterprises to deploy AI agents for common tasks—customer support, outbound sales, and high‑risk internal workflows—but the crucial detail is how it is sold. Rather than releasing another self-service API, OpenAI is routing Presence through its consulting arm, the OpenAI Deployment Company, and its Forward Deployed Engineers. Deployments are scoped individually during a limited general availability phase, led by those engineers and select systems integrators, and Presence "is not yet available as a self-serve product." OpenAI has not disclosed pricing, geographic limits, contractual terms, or the cost of the engineering and integration work that comes with a deployment. That opacity is deliberate: the company is moving into boots-on-the-ground professional services, where value is priced per outcome and complexity, not per token. In a world where base models are increasingly commoditized, Presence is OpenAI’s bet that enterprises will pay more for guaranteed agent performance, policy controls, and workflow redesign than for raw model access.
What Presence agents do differently: running the work, not sitting beside it
Presence supports realtime voice agents and chatbots that operate across customer-facing and internal workflows, with governance baked into the platform. Enterprises can define policies, permissions, and evaluation standards so agents behave within professional boundaries while answering questions, accessing company systems, taking approved actions, and escalating to human workers when needed. OpenAI’s own English-language phone support line at 1-888-GPT-0090 now runs on Presence and, according to the company, resolves 75% of inbound issues without human help, while a Codex-powered improvement loop cut human handoffs by 15 percentage points in 10 days. That is not a novelty chatbot bolted onto a help desk. A Presence agent can understand open-ended requests, verify a caller’s identity, look up account information, apply policy, and process eligible refunds—and bring in a person for exceptions. In effect, Presence aims to turn enterprise AI chatbot deployment into managed workflow automation, shrinking queues and the number of seats legacy vendors can justify.
The SaaS market impact: when the agent owns the workflow, the dashboard loses its premium
OpenAI introduced Presence on July 22, 2026, and software stocks fell as traders asked how much of the SaaS stack an AI agent can now absorb. Presence puts OpenAI directly into workflows that software vendors used to protect as their own, giving investors "a fresh reason to punish" those stocks. The overlap is obvious: Presence packages model reasoning, data access, governance, and policy control—the same capabilities many SaaS firms now rely on to defend their AI story. Presence is a managed product for putting voice and chat agents into production across customer support, outbound sales, billing, insurance claims, and internal IT requests. Every time OpenAI moves deeper into enterprise software, the key question for buyers is which existing subscription becomes easier to cancel. When an agent can read the policy, access the account, take the approved action, and escalate the edge case, the old dashboard starts to look less like the center of work and more like plumbing. SaaS companies that only "own the screen" without owning proprietary data, regulated workflow trust, or deep integrations will feel this pressure first.
What enterprises should do now: treat Presence as a test of your stack’s defensibility
Presence is available immediately in a limited general availability program, with each deployment individually scoped to the customer’s use case and implementation needs. Broader pricing details will follow as availability expands, but expectations should be clear: Presence will cost real money and is part of OpenAI’s attempt to cover massive datacenter buildout commitments and move toward profitability. One major investor has already committed USD 60 billion (approx. RM276,000,000,000) to OpenAI’s wider AI ambitions, suggesting that the company is under pressure to turn enterprise AI agents into durable revenue, not experimental tools. For enterprises already invested in other AI agent platforms, the presence of Presence raises hard questions about pricing, differentiation, and ROI: are you paying for unique capabilities, or for generic features frontier agents can replicate? The sensible move is not to chase every new agent platform, but to audit where your stack genuinely adds irreplaceable value—and where a managed agent like Presence could safely take over.
