Microsoft's AI Plan Literally Started With: "Make People Addicted"
A leaked internal Microsoft document for their new AI agent (codenamed ClawPilot, now Scout) had a phase literally titled 'Make people addicted.' Written by a Corporate VP, confirmed real by multiple outlets, and then denied by Satya Nadella himself.
So this one’s not a rumor, I checked, it’s a real leaked document, 404 Media got it first back in early June, and enough other outlets have independently confirmed pieces of it since that I’m comfortable calling this solid.
Microsoft just launched something called Scout at their Build conference this year, an “always-on personal agent” built into Microsoft 365, runs email drafting, document creation, coding tasks, the works, powered by OpenClaw under the hood. Nice normal product launch, standard productivity messaging, “human-centered AI,” all that.
Except there’s an internal strategy document, titled literally “ClawPilot: Overview and Plan with Project Lobster,” ClawPilot being what Scout was called internally before the public rename, and it lays out three phases for the rollout. Phase one, and I want to be clear this is the actual heading in the actual document, is called “Make people addicted.” Not a euphemism I’m applying. That’s the header.
The text under it says to keep shipping the standalone experience, grow the user base, and build the tool ecosystem “that makes people depend on it daily,” and then adds, and I think this is the part that actually got people upset, that this is “already happening organically.” So it’s not even aspirational at that point, they’re saying it’s already working.
Who wrote this? Not some junior product manager freelancing. The document’s credited to Omar Shahine, a Corporate Vice President at Microsoft, the same guy who wrote Scout’s official public launch blog post, alongside another exec named Jakob Werner. So this wasn’t some rogue slide deck from an intern, this came from the actual leadership running the product.
More than 1,000 Microsoft employees are apparently already using the internal version, including Satya Nadella himself, according to the document.
Here’s where it gets good though. After 404 published this, Nadella went on record with The Information and said he wasn’t sure what the document was or “who is writing and leaking this nonsense.” Which, sure, plausible deniability, except 404 came back and pointed out that Nadella knows exactly what the document is and exactly who wrote it, because one of the two names on it is the guy who wrote Scout’s own launch post, under Nadella’s own company. So either the CEO of Microsoft doesn’t read his own product team’s strategy docs before they go live at a developer conference, or he does and is just saying he doesn’t. Neither answer is great honestly.
Meanwhile Microsoft’s actual public statement on all this was, quote, “our goal isn’t more screen time, it’s more time back,” emphasizing human-centered AI and user control. Which is a genuinely strange thing to say a few days after a document with your VP’s name on it says the plan is dependency by design.
Employees inside the company were split on it too, from what 404 reported. One called the addiction language “very troubling,” said it felt like a “saying the quiet part out loud” moment. Another basically shrugged and said isn’t that the actual goal of all software from every big tech company anyway, and joked that Microsoft’s at least bad at actually pulling it off compared to companies that are better at making things addictive. Which, weirdly, might be the most honest quote in this whole story.
So here’s what I’d actually want Microsoft to answer, if anyone from PR happens to read a blog with eleven readers.
Does “make people addicted” in an internal roadmap mean something different internally than what that phrase means to literally everyone else reading it, and if so what, specifically, is that different meaning, in plain language, not a reframe.
If Nadella genuinely didn’t know about a document written by his own Corporate VP describing the flagship AI product he personally demoed at Build, what does that say about how strategy documents move through the company before a launch, and has that process changed since.
Was “already happening organically” based on actual retention data, and if so, can that data get shared, because “addiction is already happening” is either a marketing brag or an admission depending on which room you say it in, and Microsoft doesn’t get to pick both depending on who’s asking.
And genuinely, why does the public messaging say “more time back” while the internal one says “depend on it daily.” Those aren’t compatible goals. A tool that gives you time back is one you use less over time, not one designed to make you unable to stop using it.
I don’t think Microsoft invented anything new here honestly, every major AI company is fighting the same engagement-versus-wellbeing tension right now, this isn’t some uniquely evil Microsoft thing. What’s different is somebody wrote the quiet part down as an actual slide header, with their name on it, and then the CEO’s response was to act like he’d never seen it. That’s the part that’s actually damning, not that a tech company wants engagement, everyone knows that already. It’s that they said so in writing and then pretended they hadn’t.