Fewer than 7% of Microsoft's more than 450 million commercial Microsoft 365 seats worldwide currently carry a Copilot license. That single number is the real subtext behind what Microsoft announced on Friday, September 25, 2026: not a new model, but a restructuring of Copilot itself, aimed less at impressing AI researchers and more at finally getting the product in front of the roughly 93% of paying Office customers who still aren't using it.
The announcement sent Microsoft's stock up more than 4% that day — making it, per CNBC's reporting, the best-performing stock among the 'Magnificent Seven' tech giants that session, even as the broader software sector, measured by the iShares Expanded Tech-Software Sector ETF, was roughly flat to slightly lower. Here's what actually changed, and why investors reacted the way they did.
Quick summary: On September 25, 2026, Microsoft unveiled a consolidated Copilot app structure built around four areas — Home (merging the existing Chat with a new collaborative 'Cowork' mode), Code (GitHub Copilot-style development tools aimed at both engineers and non-technical employees), Office (Word, Excel, and PowerPoint functionality embedded directly), and Autopilot (a personal enterprise agent, formerly called 'Scout,' that can be assigned tasks and act inside Teams and Outlook). Pricing stays a flat $30/month per user for commercial Microsoft 365 Copilot's basic Chat tier, with the more advanced Cowork, Code, and Autopilot capabilities billed on usage. Home and Code begin rolling out first through Microsoft's Frontier early-access program 'in the coming weeks'; Autopilot expands its private preview at the end of September. Microsoft shares rose more than 4% on the news, outperforming every other 'Magnificent Seven' stock that day.
The four areas, explained
| Area | What it replaces/combines | Who it's for |
|---|---|---|
| Home | Merges existing Copilot Chat with a new 'Cowork' collaborative mode | General knowledge workers |
| Code | Brings GitHub Copilot-style coding tools into the main app | Engineers and non-technical staff attempting AI-driven coding tasks |
| Office | Embeds Word, Excel, and PowerPoint generation directly | Anyone producing documents, spreadsheets, or decks |
| Autopilot | Renamed from 'Scout'; a personal enterprise agent | Workers assigning tasks for an agent to complete inside Teams/Outlook |
Microsoft says an intelligent routing layer will eventually decide which of these four areas handles a given request automatically, so users won't have to manually pick a mode before typing — closer to how a single chat box is supposed to feel, rather than four separate apps wearing one skin.
Why 'bundling' is the actual strategy
Jacob Andreou, Microsoft's executive vice president for Copilot, framed the move in explicitly historic terms: 'every era of technology has basically had a new kind of defining interface ... with Copilot, that's really going to become this interface to AI.' The more concrete business logic is simpler: Microsoft was, by its own account, struggling to get even a tenth of its enormous installed base of paying Office 365 seats onto a separate Copilot license, and both Anthropic and OpenAI already ship bundled, all-in-one assistant experiences rather than asking users to choose between a chat mode and a coding mode. Rather than keep competing tab-by-tab, Microsoft folded everything into one app and one entry point (Source: geekwire.com).
A quiet retreat from the consumer chatbot race
One detail in the coverage is a significant strategic shift: Microsoft is stepping back from directly competing for standalone consumer chatbot market share — a race OpenAI's ChatGPT and Google's Gemini currently dominate — and concentrating instead on the enterprise seats it already has a billing relationship with through Microsoft 365. That's a bet that Microsoft's advantage isn't building the single best general-purpose chatbot, but being the AI layer already sitting inside the inbox, the spreadsheet, and the meeting people already use every day.
If your organization already pays for Microsoft 365 Copilot, the practical thing to watch isn't the rebrand — it's the usage-based pricing on Cowork, Code, and Autopilot. A flat $30/month per seat is easy to budget for; metered usage on an agent that can act inside Teams and Outlook on your behalf is not, and early adopters should track actual consumption closely before rolling Autopilot out broadly.
Reading the stock move
A single-day stock move isn't proof a product strategy will work — but a 4%+ jump that outpaced the rest of the 'Magnificent Seven' on the same day the broader software sector was slightly down is a real market signal that investors read this specifically as Microsoft addressing its weak Copilot attach rate, not as a cosmetic rebrand. Fewer than 7% of 450 million-plus seats actually using the product they're already paying for was always the headline problem with Microsoft's AI monetization story; this is Microsoft's attempt to fix the funnel rather than add another feature on top of it.
01What actually changed with Microsoft Copilot on September 25, 2026?
Microsoft consolidated its Copilot assistant into one app with four areas — Home (Chat + a new Cowork mode), Code, Office, and Autopilot (formerly 'Scout') — replacing a more fragmented set of separate tools and modes.
02Did Copilot's price change?
The base commercial Microsoft 365 Copilot Chat tier stays a flat $30/month per user. The newer Cowork, Code, and Autopilot capabilities are billed on a usage basis instead of a flat fee.
03Why did Microsoft's stock jump on the news?
Shares rose more than 4%, the best performance among 'Magnificent Seven' stocks that day, which coverage linked to Microsoft addressing a long-standing weak point — fewer than 7% of its 450 million-plus commercial Office 365 seats had a Copilot license — rather than just shipping another feature.
04How is this different from Anthropic's or OpenAI's approach?
Microsoft is explicitly positioning this as catching up to a bundled, all-in-one-app approach both Anthropic and OpenAI had already shipped — the strategic shift here is Microsoft stepping back from competing for standalone consumer chatbot share and doubling down on its existing enterprise seat base instead.
Bundling four tools into one app is one way to solve 'too many AI products, not enough clarity.' LumiChats takes a different approach to the same problem — instead of bundling one company's tools, it gives you more than 40 leading AI models from Anthropic, OpenAI, Google, and others in a single subscription, so the choice of which AI actually handles your task is yours, not pre-decided by whichever vendor you happen to be logged into.
