Microsoft has an explicit, documented goal of making users dependent on its new AI assistant, according to internal documents obtained by 404 Media. The documents, which detail the strategy behind a product then called Claw Pilot and later announced as Scout, lay out a multi-phase plan for the tool’s deployment, with phase one described simply as “make people ad*icted.”
Scout is Microsoft’s attempt to integrate OpenClaw, an open-source framework for building AI agents, directly into its Microsoft 365 suite of products. Unlike traditional chatbots that generate text a user then has to act upon, agentic AI like Scout is designed to take direct action on a user’s behalf, sending emails, managing calendars, interacting with files, and operating across Windows more broadly.
The strategy document was authored by Microsoft executive Omar Shahine and, notably, was itself generated with AI assistance. The document states it was “co-created turn by turn with AI, human verified every sentence.”
Phase one of the plan reads: “Make people ad*icted, continue shipping the standalone Claw Pilot experience, polish the UX, grow the user base, and build the skill and tool ecosystem that makes people depend on it daily. This is already happening organically.”
The word “a*dicted” in an official company document is significant. AI companies and social media platforms have faced repeated legal challenges and regulatory scrutiny over making products that foster compulsive use. Framing that goal so plainly in a written strategy document is the kind of language that tends to attract exactly that scrutiny.
According to 404 Media’s Jason Koebler, Microsoft employees who saw the document were the ones who flagged it. Some were troubled by the language on a personal level.
As Jason noted, several employees told him they had friends or family members who had struggled with dependency issues and found the explicit corporate framing offensive. Others were simply unsettled that this was the stated goal rather than, say, building a product people genuinely found useful.
The document also revealed that Claw Pilot had already been piloted internally with more than a thousand Microsoft employees, and claimed that CEO Satya Nadella was among those using it. It described the tool as having “organically grown into one of the most requested internal tools at Microsoft.”
The context matters here. Microsoft has launched several AI products in recent years that were met with significant backlash, including a Windows-integrated assistant that took periodic screenshots of user activity. Rather than building trust, those products generated public frustration. Scout represents another attempt to break through, this time with an agentic product that has far deeper access to user accounts and data than previous efforts.
The phases following the initial “make people ad*icted” goal are largely technical in nature, covering integration with Microsoft’s wider AI infrastructure and expansion into mobile and voice. Security, notably, appears as a consideration only later in the document, which several observers found telling given that an AI agent with the ability to send emails and delete files on a user’s behalf represents a meaningful security surface.
Microsoft has not publicly addressed the specific language in the document.
