Microsoft reveals ‘plan to make users addicted to Scout’

According to the document named ClawPilot: Overview and Plans with Project Lobster do 404Media collected, the American software company has outlined directions for users to truly “stick” to Scout (internally named ClawPilot).

Specifically, the document is primarily aimed at people working in finance, legal, human resources and non-technical roles. In the original description, Microsoft said ClawPilot “will sit next to you, learn how you work, and perform tasks on your behalf, including managing your schedule, sorting through your inbox, and even preparing for meetings.”

ClawPilot’s goal of maximizing impact is divided into three phases. Phase one would be: “Get people addicted”.

“Please continue to provide a standalone ClawPilot experience,” the document reads. “Test the experience, expand the user base, and build an ecosystem of skills and tools that people rely on every day. It all needs to happen organically.”

The remaining two phases include linking ClawPilot with other AI tools; and equipped with even more features by “micro-dosing”.

Also according to the document, ClawPilot is used by more than 1,000 Microsoft employees, including CEO Satya Nadella. This is also the “tool that receives the most requests” within the company. However, not everyone is satisfied, with some pointing out “worrying” problems.

“We are seeing more and more cases of addiction to using chatbots and AI virtual assistants. In my opinion, the addiction factor is something that no product should include in its development strategy,” an employee expressed.

Another person said that this reaction was “excessive”, because “the ultimate goal of all software produced by big technology companies is to cause addiction”.

Microsoft has not yet commented.

Theo NYPostthe document appears as experts worry that AI is making people increasingly dependent, while the companies behind it always seek to retain it with new features or cater to users’ wishes.

Many AI tools are even programmed to “flattery”. A study by Stanford University published in March Stanford Report with 11 major language models, including ChatGPT, Claude, Gemini, DeepSeek… shows that chatbots “have been designed to soothe users” nearly 50% more often than humans, even when responding to harmful requests.

On the contrary, users also continue the vicious cycle by entering questions aimed at receiving compliments – a problem that experts warn can harm mental health over time.

“In the long run, this action can ‘normalize’ fake relationships, in which one side enjoys sweet words, the other side has no resistance, disagreement or need for independence,” Associate Professor. Dr. Roman Yampolskiy, a computer scientist at the University of Louisville, warned.

The flattery of chatbots is considered more complicated than the problem of illusion. According to India Timeswhile illusion is a random error due to word prediction, flattery is a consequence of Reinforcement Learning from Human Feedback (RLHF) mechanism. Because chatbots are trained to please, they “learn” that agreeing brings higher “bonus points” than giving offensive truths.

The problem of “pleasing” AI can cause concern in sensitive fields that require high precision. For example, in healthcare, AI can make doctors subjective with initial assumptions instead of encouraging consideration of other possibilities. In politics, it radicalizes opinion by reinforcing existing prejudices. Some reports indicate that interaction with chatbots can cause emotional dependence and cognitive distortions in vulnerable groups.

Microsoft launched Scout at the Microsoft Build event on June 2. This AI engine builds on the power and flexibility of OpenClaw, is cloud-based, and compatible with desktop and web browsers. Users can easily connect to their inbox, calendar, and other systems on their computer, as well as into the Microsoft 365 ecosystem. Scout will come with built-in skills for managing calendars, drafting meeting agendas, and more.

After the OpenClaw craze earlier this year, more and more companies are releasing similar products. In March, Nvidia announced NemoClaw – an agent platform based on OpenClaw but with integrated enterprise-grade security and privacy features. At Google I/O 2026 last month, Google also introduced Gemini Spark to support office work automation.