As we step into 2026, one insight from the past year still echoes deeply with me.
When I first read Microsoft’s “It’s About Time: The Copilot Usage Report 2025,” I wasn’t just curious about what people ask Copilot. I wanted to understand when they ask, on which device, and what that reveals about our evolving relationship with AI.
The Microsoft AI team analyzed 37.5 million de-identified Copilot conversations from January to September 2025, focusing on two deceptively simple dimensions: time of day and device type (desktop vs. mobile). Enterprise traffic was deliberately excluded — so this study offers something rare: a window into how individuals use Copilot in their personal lives, outside corporate systems.
This article is my reflection on those patterns — and five habits I believe leaders must cultivate as AI shifts from a tool to a trusted rhythm in our daily lives.
What 37.5 Million Conversations Say About How We Use Copilot
The report shows that how people use Copilot depends fundamentally on context and device. It’s not just what they ask — it’s when and where they ask it.
1. Desktop Copilot: The Workday Colleague
Across the full dataset, the top topics overall are Technology, Work & Career, and Health & Fitness. The dominant intents are Searching, Advice, Creating, Learning, and Technical Support.
On desktop, this tilts strongly toward work:
- During 8 a.m. to 5 p.m., “Work & Career” actually overtakes “Technology” as the top topic.
- Education, science and other professional themes also rise in the workday.
On a laptop at a desk, Copilot behaves like a focused colleague – helping people gather information, write, code, edit and troubleshoot.
2. Mobile Copilot: The Constant Companion
On mobile, the pattern is very different:
- “Health & Fitness” is the most frequent topic in every single hour of the day, across all months.
- The most common topic–intent pair is Health & Fitness / Searching, and it stays at the top consistently.
Mobile Copilot becomes a constant confidant for physical well-being – the thing people turn to about their body, habits and lifestyle, no matter the time.
3. Three Rhythms: Workday, Companion, Introspective Night
Looking across the day, three clear modes emerge:
- The Workday (8 a.m. – 5 p.m.)
- The Constant Companion (all day, mobile)
- The Introspective Night
There is also a weekday vs weekend signature: in August, “Programming” ranks higher than “Games” on weekdays, but on weekends the relationship flips – “Games” overtake “Programming”. The same people who code with Copilot Monday to Friday are asking about gaming on Saturday and Sunday.
Seasonality shows up too: in February, “Personal Growth & Wellness” remains strong, while “Relationships” spikes sharply on Valentine’s Day. Copilot is quietly part of people’s emotional calendar, not just their task list.
AI won’t just be part of your stack — it will be part of your schedule.
And when you look at intents, “Searching” and “Advice” are consistently the top two. Most usage today is still augmentative – people using Copilot to explore, understand, draft and get feedback – rather than handing off full automation.
My one-line takeaway from this data:
Copilot is a productivity ally at the desk and a human companion in the pocket.
The same AI plays two very different roles depending on where and when it shows up.
My POV: Design For Context, Not Just Queries
Looking at this as someone who spends a lot of time on enterprise AI and experience design, a few reflections stand out.
1. Context is the real UI
A Copilot query at 10:00 a.m. on a work desktop is fundamentally different from a query at 11:30 p.m. on a phone.
- On desktop, the “job to be done” is usually a concrete deliverable – an email, a deck, code, an analysis.
- On mobile, especially at odd hours, the “job to be done” is often clarity, reassurance, or self-reflection.
The real UX frontier is not simplicity — it’s empathy.
Designing a single, context-blind experience misses this:
- Desktop Copilot should optimize for workflow, structure and depth.
- Mobile Copilot should optimize for brevity, empathy and safety, especially around health and personal topics.
2. We’re quietly building AI relationships
When people ask Copilot about their health, their relationships or their beliefs late at night, they are not just seeking information. They are implicitly asking:
- “Can I trust the way this AI talks to me?”
- “Does it respect my vulnerability?”
- “Will it be clear about what it doesn’t know?”
That has implications for tone, transparency, guardrails and escalation. You can’t treat those queries like another technical how-to question. As these patterns grow, duty of care becomes part of AI design.
3. Enterprises must respect both roles
Even though this dataset excludes enterprise logins, it tells us a lot about the mental model users bring to Copilot at work.
The same person who treats Copilot as a wellness or relationship companion on their phone in the evening will see it as an ally (and sometimes a critic) in their inbox and documents during the day.
Inside organizations, Copilot will increasingly be:
- An orchestration engine for tasks, documents and workflows
- A quiet coach for individuals – helping with writing, planning, learning and reflection
Governance, UX and communication should acknowledge both. If we treat Copilot purely as a “productivity add-on”, we misunderstand how people are already engaging with it.
4. Usage data must become a design input
These time-and-device patterns shouldn’t just live in a research PDF. They should directly inform:
- Safety and quality thresholds for specific topics (health, money, relationships vs casual entertainment)
- Notification and nudge strategies across time zones, weekdays and weekends
- Handoffs to humans when a conversation shifts from informational to emotional or high-risk
My own mantra after sitting with the report is:
Don’t design AI only for tasks. Design it for the moments those tasks live in.
My Top 5 Prompts or Habits to Unlock Strategic AI Value
From analyzing 37.5 million Copilot conversations, we can identify specific prompts that consistently create value—both for personal growth and professional impact. These aren’t just technical queries; they reflect how people are strategically embedding AI into the fabric of daily life. If you’re aiming to get more than just answers from AI—if you’re looking to generate insight, clarity, and direction—start with these five types of prompts:
- “Help me improve this…” Whether it’s refining a marketing pitch, tightening up a resume, or improving code, this prompt taps into AI’s core strength: augmentation. Instead of creating from scratch, users are inviting AI into an iterative, collaborative process where quality improves rapidly.
- “Summarize this and extract key insights.” Summarization remains one of the top intents, especially across work and productivity conversations. This prompt is perfect for distilling reports, articles, or meeting transcripts—turning information overload into actionable intelligence.
- “Plan this for me…” From organizing team events to mapping out product roadmaps or personal wellness goals, users often use AI for structured planning. It’s less about automation, more about structured thinking—and AI offers a logical scaffolding that saves time and mental bandwidth.
- “Compare options and recommend the best choice.” Decision support is a huge use case. This prompt works well for product research, strategy development, or even comparing career options. It activates AI’s ability to evaluate, contrast, and synthesize across variables—essential for smart decision-making.
- “Advise me on how to handle…” This is where AI starts to feel like a trusted partner. Whether it’s navigating a tough client conversation or managing a work-life balance challenge, advice-seeking prompts show that people increasingly trust AI for emotional intelligence and strategic judgment—not just facts.
From Queries to Companions: What Leaders Must Do Now
If there’s one takeaway, it’s this:
We’re not just building AI products. We’re building AI relationships.
That means:
- Building trust scaffolds, not just data pipelines.
- Designing presence-based AI, not just task completion.
- Integrating emotional intelligence into product design.
Because when users turn to Copilot at midnight to ask about philosophy — they’re not just looking for facts. They’re looking for resonance.
And if we get this right, AI becomes more than a utility. It becomes a trusted rhythm in people’s lives.
From Data to Design: Rethinking the AI Experience
These patterns of use aren’t just interesting; they are requirements for design. They push us to make AI systems that are more than just functional; they want us to make systems that are aware of their surroundings, understand time, intent, and state of mind, and are emotionally intelligent. Mobile UX must put immediacy, reassurance, and clarity first if health is the most searched topic on mobile. AI should respond to people who ask deep, philosophical questions late at night with calmness and thoughtfulness, not with transactional answers.
Leaders of businesses need to stop looking at output metrics and start looking at experience metrics: What does the AI do to make users feel during important times? Are we making systems that are useful at both 9 AM and 2 AM? Can we tell the difference between when a user wants advice and when they want information, or when they want help and when they want speed? These aren’t just technical problems; they’re also problems with how people design things. And the answers are in using both behavioral data and empathetic engineering together. This is where real innovation happens: not just in what AI can do, but in how well it can do it.
My Closing Mantra:
Design AI not for queries, but for moments. That’s where human connection lives.
Let’s build that future — not just intelligent, but emotionally intuitive.
AI’s true power isn’t just in answering questions — it’s in meeting people where they are. At midnight, during a commute, or before a pivotal moment, the best AI doesn’t just respond — it resonates. As builders and leaders, our goal should be to design systems that tune into human rhythms, not just logic. Because when AI feels present — not just smart — it earns trust. And trust is the foundation for every meaningful transformation.
Views: 4.4K
