OpenAI, Anthropic act as consultants
- Anthropic and OpenAI are no longer just selling models. In 2026, both moved deeper into enterprise rollout work through services arms, partner networks, and deployment teams. - OpenAI’s February Frontier Alliances enlisted McKinsey, BCG, Accenture, and Capgemini, while Anthropic on May 4 launched a new AI services company with Blackstone and others. - The fight is shifting from model quality to implementation — who can redesign workflows, govern risk, and actually get AI into daily operations.
Enterprise AI has hit a boring but decisive stage. The hard part is no longer getting a model demo to look impressive. The hard part is getting a giant company to change how work actually happens. That is why OpenAI and Anthropic now look a lot more like consulting firms — or at least firms that need consulting muscle — than pure model vendors. OpenAI formalized that shift in February with its Frontier Alliances, and Anthropic pushed it further on May 4 with a new AI services company built with Blackstone, Hellman & Friedman, and Goldman Sachs. ### What changed this year? OpenAI’s move was explicit. Its Frontier Alliances tie OpenAI to McKinsey, BCG, Accenture, and Capgemini in multi-year partnerships meant to help customers define strategy, integrate systems, redesign workflows, and scale deployments. Anthropic took a parallel route but with a different structure — it launched a separate enterprise AI services company that will help mid-sized firms bring Claude into core operations. ### Why does that look like consulting? Because “buy the model and good luck” has not been enough. OpenAI’s business pages now pitch expert guidance from solutions architects and an enterprise platform for “AI coworkers” connected to systems of record. That language matters. It means the sale is no longer just tokens and seats. It is process mapping, integration work, security review, and change management. OpenAI has been building the same muscle in layers. It created a Claude Partner Network, said it would invest $100 million into that ecosystem, and said it was scaling its partner-facing team fivefold with applied AI engineers and technical architects for live deals. It also struck big channel partnerships, including an Accenture tie-up and a Deloitte deployment that Anthropic called its largest enterprise rollout to date, real network. ### Why not just let Accenture do the messy part? Because the model company now needs to control the playbook. If a deployment fails, the customer often blames the model, not the systems integrator. So OpenAI and Anthropic both want a hand in the architecture, guardrails, and workflow design. Basically, they are trying to own the outcome, not just the API call. That also helps them defend enterprise accounts from each other. Axios has described OpenAI mobilizing consulting
