Why AI adoption isn't about technology
Most Copilot projects that stall don't have a technical problem. The blockage almost always comes from communication between IT teams, change management, and executive sponsors. This was the shared finding from a Microsoft webinar dedicated to the adoption of Microsoft 365 Copilot and Copilot Cowork, hosted by Karuana Gatimu (Director of the Scale Experience Group, responsible for adoption.microsoft.com) with MVP Amy Dolzine (EY, Copilot and Viva Engage MVP).
The central message: IT teams driving Copilot deployment must treat adoption as a work transformation, not as a simple software rollout. The centralized resources on adoption.microsoft.com/copilot were recently redesigned to segment content by role: user, champion, or leader.
Resource to know
The MS-4007 course covers the fundamentals of user enablement for Copilot and delivers a badge that can be showcased on LinkedIn. The Microsoft 365 Champions program provides decks and best practices each month that can be reused — no need to start from scratch.
The three pillars of an "AI-powered" organization
Three axes structure any adoption approach, regardless of organization size:
- Leadership: Executive sponsors must understand they are driving a behavioral change, not a technical deployment.
- Human change management: A distinct competency separate from training, which leverages organizational psychology and communication.
- Technical readiness: Tenant security, data governance, responsible AI principles — the foundation of trust on which everything else rests.
This foundation of trust is the variable that conditions learning. A colleague under stress cannot memorize new skills — it's a neurological mechanism, not a posture. Hence the importance of addressing AI-related anxiety before pushing features.
Business priority, not features
Not all Copilot features have the same value by role. Direct the conversation toward revenue growth and business process redesign rather than product capability lists.
Champions: the highest-leverage adoption driver
Field feedback converges: adoption speed and user satisfaction depend directly on the existence of a community of champions. A champion differs from an early adopter by their motivation: they are not there to test every novelty, but to help their colleagues.
Practical observations from organizations:
- Recruit functional experts (HR, marketing, legal) who know the business — more impactful than purely technical profiles.
- Leverage a SharePoint foundation to anchor documentation, complemented by a Teams channel or Viva Engage for community engagement.
- Provision the pre-configured Copilot adoption community in Microsoft 365, complete with ready-to-use campaigns and an agent in public preview to answer common questions.
- Recognize champions without necessarily offering monetary rewards: certificates, mentions, raffles with small incentives often suffice.
Amy Dolzine (EY) illustrates this workflow with a Power Automate flow: each training request she cannot fulfill is redirected to the champion network, accompanied by a "Copilot fundamentals" deck updated weekly — necessary as screenshots and interfaces evolve rapidly.
Common mistake with senior leaders
Never present features to a senior executive without first listening to their business problem. Amy Dolzine cites this error as the turning point in her method: leaders want results, not a feature demonstration. Plan individual or small group sessions — they don't like appearing unprepared in front of their team.
Using AI to build your own adoption plan
The webinar demonstration shows how to query Copilot directly to generate a structured adoption plan. The prompt used is intentionally simple, without change management jargon:
1I want to get people to use Copilot. We're a mid-sized consulting firm with three offices. Employees are still learning Teams, OneDrive, and SharePoint. Use adoption.microsoft.com as the primary reference to solve this problem.Elements to clarify in the prompt to refine the response:
- Industry, size (headcount or revenue), number of sites, and time zones.
- Tools already in use (complete Microsoft 365 or coexistence with Zoom, custom business applications).
- Target behavioral objective (new habits, skill development, cross-functional adoption).
In the demonstration, Work IQ was intentionally disabled to force Copilot to rely on the external content provided rather than tenant data. Result: a three-phase structure — foundations, habits, transformation — with a 90-day plan and actionable "quick wins".
Technical point to remember: Copilot remembers formatting preferences explicitly stated ("never use fonts smaller than 12 points" for example), which avoids repeating instructions with each generation.
Generated content is a starting point, not a final deliverable
Decks and plans produced by Copilot must be reworked: organization tone, branding, actual names of sites and teams. Never distribute an adoption plan without business review.
Chat, agents, or Copilot Cowork: choosing the right tool
The webinar recalls an important operational distinction between Copilot surfaces:
| Tool | Recommended use | Cost impact |
|---|---|---|
| Microsoft 365 Copilot Chat | Individual reflection, drafts, rapid iteration | Included in Copilot license |
| Pre-built agents (researcher, analyst, facilitator) | Targeted scenarios with defined agentic framework | Based on license and deployment |
| Copilot Cowork | Multi-surface tasks, autonomous, executed in background in the cloud | Consumes dedicated Copilot credits |
| Copilot Studio | Build custom agents, independence from models | Development and hosting to budget |
Practical recommendation: formulate and refine your request in Copilot Chat before transposing it to Copilot Cowork, to avoid consuming credits on a task that is still poorly defined. A complementary webinar, broadcast the day before this session, details specifically the management of Copilot Cowork credits — consult as a priority for any organization moving to this tool.
Monitoring, governance, and pitfalls to avoid
For IT and champion-side management, some concrete benchmarks:
- Measure by department, not by individual: the goal is to identify if sales adopts faster than support, not to monitor people.
- Verify necessary access rights before consulting detailed usage reports, and respect your organization's privacy constraints.
- Reuse Teams meeting transcripts as a source of truth to generate summaries and action tracking — particularly useful when transcription is only available during the meeting, depending on tenant configuration.
- Leverage the functional scenario library (filterable by product) to inspire use cases by business function, pending its full integration into the new Copilot hub.
- Consult the WorkLab site for ongoing research on AI usage at work and the concept of "frontier firm".
No perfect posture expected
Copilot evolves continuously: observed behavior today differs from months ago. Always present your demonstrations as snapshots ("here's what it did for me today"), not as a guarantee of reproducible results.
Key takeaways
- Copilot adoption rests on three inseparable pillars: leadership, human change management, technical readiness.
- Champions remain the most effective lever to accelerate adoption — recruit based on motivation to help, not technical expertise alone.
- Copilot can itself generate a personalized adoption plan: give it organizational context, existing tools, and behavioral objectives.
- Clearly distinguish Copilot Chat (exploration, drafts) from Copilot Cowork (autonomous tasks, credits consumed) to control costs.
- Measure adoption at the department level, train executives individually, and celebrate champions without expecting monetary rewards.
The next concrete step for any IT team: provision the pre-configured Copilot adoption community in the tenant, then plan a first monthly cycle of champions before expanding deployment to Copilot Cowork.
