Microsoft’s Sales Team Reportedly Gets a New Target: Its Own AI Partners in OpenAI and Anthropic

Microsoft's Sales Team Reportedly Gets a New Target: Its Own AI Partners in OpenAI and Anthropic

Microsoft’s sales team reportedly has a new pitch for the financial year ahead, one that takes direct aim at Anthropic, OpenAI, and Google. Naturally, these are the same companies whose models have, until recently, powered parts of Microsoft’s own AI products.

Microsoft’s Pitch

According to a Bloomberg report, Microsoft executives used an internal strategy session on 15 July to brief salespeople on how to position the company’s in-house models against its rivals. In places, this included its own partners. Executive Vice President Jay Parikh reportedly set the tone. He argued that Microsoft’s edge lies in breadth rather than any single component.

“Everyone else is selling parts — we’re selling the full end-to-end system.”

Copilot Executive Vice President Jacob Andreou then went after Anthropic’s Claude specifically, comparing it directly with Copilot’s performance inside Microsoft’s office apps. “Slower and less accurate, and lacked the proper security integrations,” he said.

Neither Microsoft nor Anthropic had responded to Bloomberg’s request for comment at the time of the original report.

This isn’t an isolated jab. A separate Bloomberg report earlier this month found Microsoft has been swapping OpenAI and Anthropic models out of Word and Excel in favour of its own MAI-branded versions. This is a change pitched as cost-saving rather than performance-driven.

Meanwhile, back in April, Microsoft and OpenAI dropped the exclusivity clause that once tied OpenAI’s models to Microsoft’s products. This freed OpenAI to sell to Microsoft’s competitors. Put side by side, three moves in three months plausibly sketch a company rewriting its own supply chain.

Reading the Room With the AI Giants

None of this happens in a vacuum. Microsoft’s share price has had a rocky stretch this year. Investors have been pressing the company on the scale of its AI infrastructure spend without an obvious payback story yet. Standing up and declaring your own models the strongest in the room is a straightforward way to at least try to course-correct that narrative. At least for a quarter, that is.

The story breaks at a problematic moment for the wider market. CX leaders and tech buyers are already nervous about betting too much on one AI vendor. Recent surveys of technology leaders put the share of running several models in parallel, rather than standardising on one, well above a third and climbing. A supplier getting into it with its own suppliers, to some extent, publicly isn’t going to settle those nerves.

What it Means for CX and IT Leaders Around Vendor Lock-In  

If your organisation runs Microsoft 365 Copilot, the potential issue is whether you actually know which model sits behind the features your teams use every day, and whether that’s changed in the past few months without anyone mentioning it.

Copilot passed 20 million paid seats earlier this year, so this touches a lot of desks. It’s maybe worth an informal conversation with your account team. Topics could include what’s powering the tools now, how it has shifted recently, and whether your contract says anything about being told if it happens again.