Human beings are difficult to manage. Managing AI agents is… hard, too. That’s why OpenAI is launching a new platform called OpenAI Frontier, which it says will help businesses “build, deploy, and manage” AI agents, even those not created by OpenAI itself.
Frontier’s description of OpenAI sounds like HR for AI. “Frontier gives agents the same skills people need to succeed at work: shared context, onboarding, hands-on learning with feedback, and clear permissions and limits,” OpenAI wrote in a blog post. The similarity makes sense: OpenAI said the product was “inspired by looking at how enterprises already measure people.”
The Frontier is available today, though only to an unspecified “limited group of customers, with broader availability in the next few months.” OpenAI said Intuit, State Farm, Thermo Fisher and Uber are among the first companies to adopt OpenAI Frontier, with “dozens of existing customers” also piloting it. It’s also unclear how much the Frontier will cost. In a press conference, Chief Revenue Officer Dennis Dresser declined to disclose pricing at this time.
Frontier is an “agent interface,” said Barrett Zoff, OpenAI’s general manager of business-to-business, who recently returned to OpenAI after a stint at the Thinking Machines Lab. Right now, many companies run AI agents on top of whatever they’re using, which often means fragmented devices, disconnected workflows, and siled data. Frontier sits on top of creating a “shared business context” for agents, connecting them to everything they need to work and communicate effectively. These connections mean deployed agents can operate in a variety of environments, though OpenAI said Frontier also lets users set limitations, making it “possible to use them with confidence in sensitive and regulated environments.”
Frontier will also make it easier for human teams in organizations to “hire AI colleagues” for tasks like running code and analyzing data. OpenAI said the agents “will also form memories” and can be evaluated by human workers, which should make them more useful over time.
OpenAI’s ultimate goal sounds suspiciously similar to Sauron’s inspiration lord of the rings: One platform to rule them all. “By the end of the year, the majority of digital work in major enterprises will be directed by people and executed by fleets of agents,” said Fidzi Simo, CEO of Applications at OpenAI. “And the dream I had was to have one platform to create and manage all of our agents.”
Interestingly, it means “a recognition that we’re not going to build everything ourselves,” Simo said. Frontier will use open standards and can be populated with agents built by OpenAI, enterprise customers, or any other AI company.
The frontier comes as AI companies attempt to prove that AI tools are actually useful to their customers, and work to create revenue streams that justify the huge amounts of money being poured into this area. Agents, devices that can act independently, are the main focus for this, and Frontier can be seen as a direct response to Microsoft’s Agent 365 agent manager. Anthropic also has stiff competition, with its cloud cowork and cloud code taking the AI industry by storm.
