OpenAI launched Frontier, an agentic platform for automated workflows in areas like customer support and IT operations, on February 5 — a direct response to Anthropic's open-source plugin push — then followed on February 26 with Frontier Alliances, a program partnering with major consulting firms to drive enterprise deployment.
The platform
Frontier packages OpenAI's models into deployable automation for the two enterprise workflows with the clearest AI ROI story: support ticket resolution and IT operations. The strategic logic is straightforward — as model capabilities converge at the frontier, the labs are racing to own the layer where work actually gets automated, because that is where switching costs live.
The consultants
Frontier Alliances is the more telling move. Enterprise AI deployments stall in the last mile — integration, change management, governance — and that last mile is precisely what consultancies sell. By formalizing partnerships with large consulting firms, OpenAI is buying implementation capacity it cannot build fast enough internally, and putting its models at the front of the recommendation queue when Fortune 500 boards ask their advisors what to deploy.
The subtext
February also saw OpenAI hire Peter Steinberger, creator of the viral OpenClaw assistant, to lead what Sam Altman described as the next generation of personal agents. Platform, distribution, talent: the through-line of OpenAI's February was that the model is no longer the product. The deployment is.