OpenAI Presence: The Trusted AI Revolutionizing Businesses

OpenAI Presence: A Leap Forward in AI Agent Reliability

In the business world, the challenge is no longer to demonstrate that artificial intelligence agents can function, but to make them reliable enough to perform high-value tasks. These agents must be capable of adapting to changes in products, policies, and user behaviors. This goes far beyond merely using an AI model: it involves establishing robust systems, rigorous evaluations, and deployment expertise to continuously improve the agents without losing control.

The Introduction of OpenAI Presence

OpenAI recently launched Presence, a product designed to help businesses deploy trustworthy AI agents. These agents can answer questions, solve problems, interact with company systems, take approved actions, and call on humans when necessary. Drawing on several years of collaboration with large-scale clients, Presence combines the reasoning of AI models with policies, safeguards, and escalation rules to ensure the accuracy and performance of the agents.

Each deployment begins with a specific task, such as resolving billing issues or managing IT service requests. The agent receives only the knowledge and access necessary to accomplish this task. Companies define the policies that determine what the agent can do, when it should seek approval, and when human intervention is required. After launch, production sessions and escalations help identify gaps. Codex then proposes updates that teams can test and approve, helping the agent adapt to changes in customer behavior.

Personalized Support for Every Business

OpenAI works closely with each client to identify high-value workflows, connect systems, and establish the necessary permissions and policies. The agent is then tested and put into production. As the deployment expands, OpenAI and certain system integrators continue to provide support. Presence was developed in close collaboration with OpenAI's research team, and each deployment generates insights that feed ongoing research and product development, thus enhancing the experience for all clients.

Availability for Voice and Chat Agents

Today, Presence is capable of handling real-time interactions via voice and chat, covering areas such as customer support, outbound sales, and high-risk internal workflows. A concrete example is resolving billing issues, where the agent understands the request, verifies the customer, searches for account information, applies company policy, and takes approved actions.

Businesses can determine what should remain constant across deployments, such as policies and escalation rules, and what should be adjusted for each workflow or channel. This allows teams to build on what works and expand to new use cases without starting from scratch.

Presence encompasses all the necessary components to operate agents in production: policies and standard operating procedures, safeguards, approved actions, simulations, assessment tools, and an improvement process powered by Codex. Together, these elements help teams connect company systems, define agent behavior, evaluate performance, enforce policies, and manage changes post-launch.

A Proven Solution in Large Enterprises

Presence has been shaped by years of deploying agents with clients, meeting the demands of critical environments. Each deployment has generated insights that continuously inform OpenAI's research and product development, creating cumulative benefits for clients.

For example, Presence powers OpenAI's English-language support hotline at 1-888-GPT‑0090, managing open requests, verifying callers, using account context, and taking approved actions. Within weeks, it met or exceeded the quality criteria of human frontline support, now resolving 75% of incoming issues without human assistance. Thanks to Codex, the improvement cycle has reduced human transfers by 15 percentage points in just 10 days.

Partner Testimonials

Renowned companies are also relying on this proven foundation. BBVA is exploring AI-powered voice support for everyday banking needs in Mexico. SoftBank is testing natural Japanese customer conversations. IAG is exploring timely support during high-demand events such as severe weather conditions.

• Daniel Ordaz, Head of AI Transformation at BBVA Mexico, states: "At BBVA, we are working closely with OpenAI to explore how trustworthy customer agents can shape the future of financial services. As a design partner for Presence, we are collaborating with OpenAI to refine voice experiences for financial customer service, as part of our goal to provide a faster, smoother, and more personalized experience with every interaction."

• Tadahisa Murakami, Vice President, Head of Data and Digital Transformation Division at SoftBank Corp., adds: "Through our collaboration with OpenAI, we are exploring how Presence can enable trustworthy customer agents that communicate naturally, connect to the necessary processes to resolve requests, and represent SoftBank consistently across customer interactions. Our frontline teams have rated the agent's conversations in Japanese very positively for their natural and accurate quality."

• Julie Batch, CEO of Retail Insurance Australia, IAG, comments: "We see an opportunity to use Presence to deploy trustworthy customer agents that provide timely and reliable support to our customers during critical moments, especially during high-demand periods such as severe weather events and natural disasters. Our goal is to build this capability in a way that helps customers navigate the claims process, strengthens our service model, and ensures that support is readily available when customers need it most."

Built for Trust Before, During, and After Launch

Before a Presence deployment reaches users, teams can test it against common requests, edge cases, and higher-risk scenarios. Simulations and evaluators check whether the agent achieves the right outcome, follows policy, uses tools correctly, and escalates when appropriate. Safeguards can intervene when the interaction goes beyond company limits.

This work does not stop at launch. Production sessions, escalations, and quality signals show teams where an agent is performing well and where it needs attention as policies, products, and user behavior change. Codex, using the Presence plugin, investigates these signals and suggests updates. Teams can test each proposed change against the production version and then approve a controlled deployment.

Presence is also designed to evolve with businesses, improving as customer and employee needs change while maintaining control. Based on OpenAI's cutting-edge research, it continues to advance as models and capabilities improve. When a use case exceeds what the product currently supports, OpenAI's Deployed Engineers and partners can work with the client to bring it into production.

Availability

OpenAI Presence is available for eligible enterprise clients through a limited general availability program. Deployments are led by OpenAI's Deployed Engineers and certain global system integrators. Presence is not yet available as a self-service product.

As we introduce OpenAI Presence, we will continue to support voice clients with access to our state-of-the-art models via the OpenAI API. To explore whether Presence is suitable for your organization, please contact your OpenAI account team.