The New Segmentation Strategy: Not Personas. Human Rhythms.

People do not use AI in one way.
They use it in modes that track their energy, uncertainty, and attention.
The data in Microsoft’s Copilot report makes that impossible to ignore.

1. Core Findings

The study analyzed 37.5 million conversations and the pattern is clear:
AI behavior is not explained by personas. It is explained by context.

• •

  Mobile acts like a confidant.
  Health and self-regulation dominate mobile queries. The device sits closest to anxiety, so the agent becomes a first responder for routines, reassurance, and personal stability.

• •

  Desktop acts like a colleague.
  Workday usage clusters around programming, analysis, planning, and structured execution. The agent becomes a cognitive force multiplier.

• •

  AI follows human rhythms, not product categories.
  Weekday coding. Weekend gaming. Late-night philosophy. Commute-time travel planning. The agent shadows human tempo, not feature sets.

Underneath this is one mechanism.
People use AI where cognitive load spikes faster than their available support.

2. Product Strategy Implications

Most builders still ship one Copilot with one tone and one capability set. The data breaks that model.

• •

  Surface-specific personalities become necessary.
  Desktop requires an agent built for density, planning, and multi-step workflows.
  Mobile requires an agent tuned for brevity, grounding, and emotional load.
  Single-pattern chatbots cannot adapt to these differences.

• •

  Advice-seeking now rivals fact-seeking.
  Users are shifting from information retrieval to decision scaffolding. The center of gravity is moving toward guidance.

• •

  Trust becomes a functional requirement.
  When users bring health, relationships, and self-worth into the conversation, tone and safety become part of the product skeleton.

• •

  Context becomes the new UX.
  Time, device, and cognitive mode should automatically shape agent behavior. Static prompts cannot do this. Adaptive agents can.

3. User Base Evolution

The report maps a clear adoption arc across 2025.

• •

  From technical users to the mainstream.
  Early usage was developer-heavy. By September, general-purpose and personal queries dominate.
  This shift happened in nine months.

• •

  From exploration to intent.
  The early noise of “Hi” messages fades. Users learn how to command the system with purpose.

• •

  From doing to being.
  Late-night spikes in reflective and existential queries show the shift.
  AI is no longer just a productivity tool. It is becoming the place where people think.

The Copilot data is a quiet warning.
The next generation of AI products will not win by being smarter.
They will win by meeting users in the mode they are already in.

Source: https://microsoft.ai/wp-content/uploads/2025/12/What_people_do_with_Copilot-8.pdf