According to PCWorld, Microsoft’s Copilot AI assistant is facing a crisis of relevance among power users, who are actively removing it from their taskbars in favor of other chatbots like OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude. The article draws a direct comparison to the era when users skipped the built-in Internet Explorer for superior browsers like Firefox and Chrome. It notes that while Copilot routes requests to OpenAI’s GPT models, the ChatGPT platform itself is more mature and configurable. The piece cites an internal email from Microsoft CEO Satya Nadella, reported by The Information, where he admitted Copilot’s connections to Gmail and Outlook largely “don’t really work” and are “not smart.” Furthermore, The Verge’s Tom Warren reported that Anthropic’s Claude Code is increasingly favored internally at Microsoft over GitHub Copilot.
The IE Parallel is Spot-On
And you know what? This comparison feels painfully accurate. Internet Explorer wasn’t terrible because it was technically incapable; it was terrible because it was the complacent, pre-installed option that didn’t have to try. Microsoft is in serious danger of replicating that exact dynamic with AI. They have this massive installed base via Windows and Microsoft 365, so Copilot gets that coveted taskbar spot. But if the experience isn’t best-in-class, that advantage becomes a liability. It just reminds people there’s something better out there. So they go find it.
Where Copilot Actually Wins (And Loses)
Here’s the thing: Copilot’s real strengths are in very specific, locked-down environments. GitHub Copilot for developers is a genuine market leader. And the deep integration of Microsoft 365 Copilot with an organization’s own data in Word, Excel, and Outlook is its killer app for enterprises. But that’s the problem. The free, consumer-facing Copilot that lives on your taskbar? It doesn’t get those benefits. It’s a watered-down version. The PCWorld story about the Claude user who grudgingly used Copilot in Word to make a chart says it all. It did the job, but the intelligence felt inferior. When your flagship AI makes people appreciate the competition more, you’ve got a fundamental product issue.
The Internal Doubt is Deafening
But the most damning evidence isn’t from users. It’s from Microsoft itself. When your own CEO sends an email saying key features “don’t really work,” that’s a five-alarm fire. And if internal teams are reportedly preferring a competitor’s coding tool (Claude Code) over your own (GitHub Copilot), what does that say about the product’s technical edge? It creates a really bad look. If Microsoft doesn’t even prefer Copilot, why should anyone else? It completely undermines confidence.
Can Microsoft Fix This?
So, what’s the path forward? They can’t just rely on bundling. The IE strategy of riding on Windows monopoly power doesn’t work in the cloud AI era. Users will simply bypass it. Microsoft needs to make the base Copilot experience so good that people want to use it even without the Office integration. They need to match the raw capability and conversational feel of ChatGPT and Claude. The desktop vision feature is a cool start, but it’s not enough. Basically, they have to compete on merit, not just on default settings. Otherwise, the “new Internet Explorer” label is going to stick. And in the fast-moving AI race, that’s a recipe for irrelevance.
