Sam Altman, CEO of OpenAI, is pictured on September 25, 2025 in Berlin.
Florian Gaertner | Photo library | Getty Images
OpenAI announced a new enterprise platform called Frontier on Thursday, its latest launch as part of its ongoing efforts to win more business customers.
Frontier acts as an intelligence layer that brings together disparate systems and data within an organization. OpenAI said the platform will make it easier to manage, deploy and build businesses. artificial intelligence agents, which are tools capable of performing tasks independently on behalf of a user.
“Frontier is really a recognition that we’re not going to build everything ourselves,” Fidji Simo, OpenAI’s applications CEO, told reporters at a press briefing. “We will work with the ecosystem to build alongside it, and we recognize that businesses will need many different partners.”
OpenAI has made an aggressive push into enterprises in recent years and announced in November that more than 1 million business customers around the world use the company’s technology.
Last month, OpenAI CFO Sarah Friar told CNBC that enterprise customers account for about 40% of OpenAI’s business, although it expects that figure to approach 50% by the end of the year.
OpenAI’s new Frontier platform is “complementary” to its existing commercial offerings, including ChatGPT Enterprise, the company said.
“What’s really missing, for most companies, is just a simple way to unlock the power of agents as teammates, able to operate within the company without needing to rework everything underneath.” Denise CommodeOpenAI’s chief revenue officer said during the briefing. “That’s exactly why we built Frontier.”
OpenAI declined to share platform pricing details.
Frontier is compatible with agents created by OpenAI, agents that companies create themselves, as well as third-party agents like Google, Microsoft And Anthropic. Simo said it’s simply not possible for OpenAI to “create all the AI agents that businesses need.”
The platform allows agents to access a “shared business context” within an organization by connecting siled internal applications, ticketing tools and data warehouses, OpenAI said.
In this context, AI agents will be able to handle complex tasks and reason about data in an open agent execution environment, according to a blog post. This means that employees of a company can ask agents to use tools on a computer, run code and work with files, for example.
Frontier also includes built-in tools to evaluate and optimize agent performance, which should help them improve over time, OpenAI said.
“What we’re basically doing is turning agents into real AI colleagues,” Barret Zoph, OpenAI’s general manager of business-to-business, told reporters.
Zoph joined OpenAI in January after abruptly leaving Thinking Machines Lab, an AI startup he co-founded alongside former OpenAI CTO Mira Murati.
OpenAI said Frontier was initially launching to a small group of customers. Early adopters include organizations like UberState farm, Intuition And Thermo Fisher Scientific. The company said wider availability would arrive in the coming months.
WATCH: Watch the full CNBC interview with OpenAI CFO Sarah Friar