OpenAI introduces Presence
On July 22, 2026, OpenAI introduced OpenAI Presence, an enterprise product designed to help organizations move AI agents from experimental prototypes into real production workflows.
Rather than functioning only as conversational assistants, Presence agents can answer questions, access relevant company knowledge and systems, take approved actions, and escalate interactions to human employees when necessary.
OpenAI argues that the challenge for enterprises is increasingly not whether AI agents can perform useful work, but whether they can perform high-value work reliably while remaining under organizational control.
Agents are built around specific jobs
Each Presence deployment begins with a defined task, such as resolving billing issues, supporting insurance claims, or handling employee IT service requests.
The agent receives only the knowledge and system access needed for that job.
Organizations define the policies governing what an agent can do, which actions require approval, and when a human employee should take over.
This approach is particularly relevant for organizations introducing agents into higher-risk business processes.
Real-time voice and chat agents
Presence currently supports real-time voice and chat experiences.
OpenAI lists customer support, outbound sales, and higher-risk internal workflows among its early use cases.
A support agent, for example, could understand a billing request, verify the customer, retrieve account information, apply company policy, and take an approved action.
This moves the role of AI beyond answering questions and toward actively participating in enterprise workflows.
Policies, guardrails and human escalation
Presence combines several components needed to operate production agents, including policies, standard operating procedures, guardrails, approved actions, simulations and evaluation tools.
Before deployment, teams can test agents against common requests, edge cases and higher-risk scenarios.
Evaluations can determine whether an agent reached the correct outcome, followed company policy, used tools appropriately and escalated the interaction when required.
Codex powers the improvement loop
Another notable part of Presence is its integration with Codex.
After an agent enters production, real sessions, escalations and quality signals can reveal weaknesses or changes in customer behavior.
Codex, using the Presence plugin, can investigate those signals and propose updates.
Teams can then test proposed changes against the production version and explicitly approve a controlled rollout rather than allowing modifications to reach production automatically.
This gives organizations a mechanism for continuously improving agents while maintaining operational control.
OpenAI is using Presence for its own support operations
OpenAI says Presence powers its English-language phone support channel.
The agent can handle open-ended requests, verify callers, use account context and take approved actions.
According to OpenAI, the system now resolves 75% of inbound issues without human assistance.
The company also reports that its Codex-powered improvement loop reduced human handoffs by 15 percentage points within 10 days.
Enterprises are already exploring Presence
OpenAI identified several major organizations working with the technology.
BBVA is exploring AI-powered voice support for everyday banking interactions in Mexico.
SoftBank is testing natural Japanese-language customer conversations.
IAG is exploring how agents could provide timely support during high-demand events such as severe weather and natural disasters.
These deployments demonstrate that Presence is positioned as operational agent infrastructure rather than merely another chatbot-building product.
Why Presence matters for Agentic CRM
Presence is particularly relevant to the emerging concept of Agentic CRM.
Traditional CRM systems primarily store customer information and provide tools for human users. Agentic CRM architectures introduce autonomous or semi-autonomous agents that can participate directly in customer and operational workflows.
A customer-service agent could, for example, retrieve customer context from the CRM, apply company policies, invoke an approved API, perform an action, update the CRM record and escalate the case if risk or uncertainty becomes too high.
Presence focuses on many of the infrastructure requirements needed to operate this type of system, including permissions, evaluations, guardrails, actions and human escalation.
The announcement therefore reflects a broader industry shift from AI assistants that primarily recommend actions toward operational AI agents that can execute approved actions within clearly defined organizational boundaries.
Availability
OpenAI Presence is currently available to eligible enterprise customers through a limited general availability program.
Deployments are led by OpenAI Forward Deployed Engineers and selected global systems integrators.
Presence is not currently offered as a self-service product.
OpenAI also says it will continue providing voice customers access to its frontier models through the OpenAI API.
Source
This article was independently summarized and rewritten based on OpenAI's official Introducing OpenAI Presence announcement published on July 22, 2026.
Read the official OpenAI announcement
This is an independent news summary and analysis rather than a direct reproduction or translation of OpenAI's announcement. Sections discussing Agentic CRM contain additional editorial analysis.