Internal presentation by Microsoft's CFO Amy Hood: number of users for Copilot has been flat over the past year with ~20M weekly users, as ChatGPT grew to 400M
It's about as useless as their search bar. Linux is looking more and more tempting every day.. 😅 [embedded post] Mastodon: Dare Obasanjo / @carnage4life@mas.to : Microsoft acqui-hired Mustafa Suleyman's Inflection AI for $650M as a back up plan to compete against OpenAI's models and ChatGPT. — So far it seems none of this has panned out. CoPilot is stuck at 20M active users while ChatGPT has 400M. And no one talks about Microsoft's AI models. … X: Tom Dotan / @cityofthetown : Microsoft's consumer AI tool Copilot has been flat for a year and a fraction the size of ChatGPT, according to an internal presentation by CFO Amy Hood. One year is also the length of time Mustafa Suleyman has been leading the effort to turn things around https://www.newcomer.co/... Barry Schwartz / @rustybrick : Tough title but true? [image] Forums: Msmash / Slashdot : Microsoft's Big AI Hire Can't Match OpenAI
Context & Ripple Effects
Microsoft had already positioned Copilot across work and consumer surfaces, first by bringing GPT-4 into Office through Microsoft 365 Copilot and then launching a free Copilot mobile app with GPT-4 and DALL-E 3. The internal usage figure tests whether that distribution translated into an independent consumer habit.
The comparison with ChatGPT matters because it separates product availability from sustained audience reach: Copilot is present in Microsoft's ecosystem, but the presentation indicates that reach has not converted into comparable weekly use.
First-order effects
- The flat roughly 20M weekly-user figure makes consumer Copilot's adoption gap with ChatGPT visible inside Microsoft, increasing pressure on the team led by Mustafa Suleyman to improve engagement rather than simply expand availability.
- Microsoft's consumer AI strategy must now be judged against repeat usage, not the earlier rollout of Copilot across mobile and productivity products.
Second-order effects
- The gap weakens the assumption that Microsoft's existing software distribution alone can win consumer assistant usage; rivals can compete on the destination product and user habit, not only on platform access.
- Microsoft may need to put greater weight on distinct Copilot experiences and retention if it wants its consumer tool to become a meaningful counterpart to its workplace-oriented AI offerings.
Third-order effects
- If assistant use continues to concentrate around a small number of consumer destinations, AI distribution advantage will depend less on bundling and more on becoming the user's default work surface.
- The divergence could reinforce a split between broadly used consumer assistants and AI embedded in existing enterprise software, unless integrated products demonstrate a reason for users to return.
The trend: This is one data point in the shift from AI launch breadth to the harder contest for recurring assistant usage and default-user status.