Microsoft reportedly trains salespeople to talk down OpenAI and Anthropic

A TechCrunch report says Microsoft is training its sales force to actively disparage OpenAI and Anthropic when pitching Copilot — a striking posture given Microsoft's deep, multibillion-dollar partnership with OpenAI. Microsoft AI EVP Jacob Andreou reportedly compares Copilot head-to-head with Anthropic's Claude, characterizing Anthropic's model as slower, less accurate, and lacking the security and governance integrations Copilot enjoys natively inside Office and Microsoft 365.

The messaging reveals the awkward reality of Microsoft's AI stack: it resells OpenAI models across M365 Copilot, GitHub Copilot, Word, Excel and PowerPoint, yet increasingly positions Copilot as its own differentiated product, and it competes with Anthropic in the enterprise-assistant market. Training reps to knock rivals suggests Microsoft sees Claude in particular as a genuine enterprise threat now that Anthropic is pushing into healthcare (Optum, UST) and security.

The backdrop is a broader vendor-rivalry moment: Sam Altman spent a day publicly roasting Anthropic on X, and Anthropic told the U.S. Senate that Alibaba ran 25,000 fake accounts and 28.8M Claude conversations to copy the model. The gloves are off across the industry. Separately, Satya Nadella warned of AI risks and signaled a new regulatory era, and Mustafa Suleyman touted a Nature Health study on Copilot's health value — Microsoft working both the competitive and trust angles.

Skeptics note that competitive sales talk tracks are standard in enterprise software, and that a leaked playbook doesn't equal official strategy. What to watch: whether the OpenAI relationship strains publicly, and whether Anthropic responds to the characterization of Claude as slower and less secure.