Study guide
11.3 Establish an AI Champions Program
Microsoft Certified: AI Transformation Leader Exam AB-731
Track your progress as you read
Finished sections get a check mark.
1 Introduction & Exam Overview
0.1 Current AB-731 Exam Facts 0.2 Audience Profile and the July 22, 2026 Blueprint 0.3 Independent Study Path for AB-731
2 Generative AI Foundations
1.1 Generative AI Versus Other Types of AI 1.2 Select a Generative AI Solution for a Business Need 1.3 Pretrained Versus Fine-Tuned Models
3 Cost, Risks, and Business Value
2.1 Tokens and ROI Cost Drivers 2.2 Fabrications, Reliability, and Bias 2.3 Scalability, Automation, and When Generative AI Adds Value
4 Prompt Engineering
3.1 Impact of Prompt Engineering 3.2 Prompt Engineering Techniques
5 Grounding, RAG, and Data
4.1 Business Requirements for Grounding Solutions 4.2 Retrieval-Augmented Generation (RAG) 4.3 Data Type, Quality, and Representative Datasets
6 Secure AI and Machine Learning
5.1 Importance of Secure AI 5.2 Application, Data, and Authentication Security for AI 5.3 When Machine Learning Adds Value 5.4 Lifecycle of a Machine Learning Solution
7 Microsoft 365 Copilot and Microsoft Copilot
6.1 Map Business Processes and Use Cases to Copilot 6.2 Differences in Capabilities Between Versions of Copilot 6.3 Microsoft 365 Copilot Chat on Web and Mobile 6.4 Copilot in Microsoft 365 Apps
8 Copilot Studio, Microsoft Graph, and Integrated AI
7.1 Microsoft Copilot Studio Capabilities 7.2 Microsoft Graph Capabilities 7.3 Integrated Microsoft AI, Risk Mitigation, and Safety
9 Agents, Extensibility, and Stack Mapping
8.1 Map Processes to Microsoft AI Apps and Services 8.2 When to Use Researcher or Analyst in Copilot 8.3 Build, Buy, or Extend Copilot
10 Microsoft Foundry and Foundry Tools
9.1 Map Business Processes to Foundry Tools 9.2 Azure Vision, Azure AI Search, and Microsoft Foundry 9.3 Match an AI Model to a Business Need 9.4 Scalability and Security Benefits of Foundry
11 Responsible AI and Governance
10.1 Importance of Responsible AI 10.2 Governance Principles for AI Use 10.3 Establish an AI Council 10.4 Microsoft Responsible AI Standards
12 Adoption Teams, Barriers, and Champions
11.1 Establish an Adoption Team 11.2 Common Barriers to AI Adoption 11.3 Establish an AI Champions Program
13 Impacts, Copilot Licensing, and Foundry Cost Models
12.1 Impacts to Data, Security, Privacy, and Cost 12.2 Copilot License Types 12.3 Foundry Tools Subscription Models
11.3 Establish an AI Champions Program
Key Takeaways Explain This Topic AI
Microsoft's Copilot playbook tells leaders to identify champions and early adopters from the AI council, the user community, manager nominations, and Microsoft Copilot Dashboard usage—not from whoever happens to sit nearest IT.
Microsoft Adoption's Microsoft 365 Champion Program describes champions as peer influencers who spark enthusiasm, surface business problems, and feed the project team; they are not a hidden overflow helpdesk.
IT keeps identity, DLP, license assignment, and tenant incidents; champions coach prompts and workflows, host office hours, and carry friction to the adoption team and council.
A healthy program includes selection criteria, manager-protected time, a community of practice (Teams and Viva Engage in Microsoft's playbook), and a Forms-style feedback loop into council meetings.
Champion health is measured by coaching delivered, community activity, protected time actually taken, signal quality to the council, and turnover—not by how many severity-1 tickets champions closed.
Last updated: September 2026
11.3 Establish an AI Champions Program
Quick Answer: Champions are peer coaches who show Copilot value in real work, run office hours, and send friction upstream. They are not tenant admins and not a 24/7 helpdesk. Select them on influence plus practice, protect their time, give them a community of practice, and route their signal to the AI council. Measure coaching and community health, not tickets dumped on volunteers.
Microsoft's July 22, 2026 skills list includes Establish an AI champions program as its own adoption bullet. The skill is easy to fake: publish a Teams emoji, call twelve enthusiasts "champions," and hope. AB-731 expects a leader to know why that fake program dies, and what a real one requires.
What Microsoft actually publishes
Stay with public pages. Do not import unofficial internal nicknames as if they were products every customer must stand up.
Playbook Step 5 — Create a Copilot user community (peer learning).
Playbook Step 6 — Identify Copilot champions and early adopters .
Playbook Step 10 — Publicly celebrate successes , including power users.
Become a Champion on Microsoft Adoption — the public Microsoft 365 Champion Program .
Learn's adoption unit — concrete community tools: a Teams space, Viva Engage , Microsoft Forms , and the Microsoft Copilot Dashboard .
Microsoft Adoption states that champions combine technical acumen with people skills. They evangelize and help their peers . The same page lists champion skills that map cleanly to a workplace program:
Create enthusiasm that increases adoption.
Build a circle of influence on their teams.
Bring new ways of working to life.
Identify business challenges and possible solutions.
Provide feedback to the project team and sponsors .
Reduce strain on the core project team through ongoing engagement.
That last bullet is the trap. "Reduce strain" does not mean "route every password reset to the nearest champion." It means the core adoption team is not the only human who can show a prompt in Word. Strain reduction is peer coaching . Helpdesk dump is unpaid IT .
Microsoft's playbook says champions spark interest and show real benefits in daily work . Learn's unit adds a practical coaching habit: when champions share tips, lead with a concrete prompt and the value it produced ("I saved three hours drafting this sales script in Word using this prompt"), not a feature tour. You may use that method without copying any internal Microsoft Digital community brand name.
Champions versus IT versus the helpdesk
Draw this line in the launch FAQ. If you do not, managers will fill it in for you.
- Work Champions IT / Copilot admin Service desk
- Show a role-specific prompt in Word, Excel, Teams, Outlook, PowerPoint Yes Rarely No
- Host office hours and walk a teammate through Copilot Chat Yes Optional guest No
Collect "this answer cited the wrong policy" patterns Yes, then escalate Receives the pattern Tickets only if it is an outage
Assign or reclaim Copilot licenses No Yes No
Identity, multifactor authentication (MFA) , DLP, sensitivity labels, oversharing No Yes Password and access tickets
- Approve a Copilot Studio agent for production Input on the business need Control plane, access, analytics No
- Sev-1 tenant outage Communicate workaround if asked Own the incident Own the ticket
- Responsible-AI policy and risk appetite Model the policy in coaching Technical enforcement No
A legal director who asks a champion to "just turn off Copilot for the whole department" is asking the wrong person. That is an IT and council decision. A champion who spends Friday nights closing hardware tickets will quit, and the program becomes a cautionary slide.
Selection: influence plus practice, not job title
Microsoft's Learn unit lists four ways to find Copilot champions and early adopters:
Lean on the AI council to name power users.
Watch the user community for people who already share tips or feedback.
Ask managers for a team representative.
Use the Copilot Dashboard to measure usage.
Build a simple score, then interview. You want people others already ask for help, who will follow data rules, and who can tell a story in the team's dialect. You do not want only the Copilot-obsessed analyst in a corner, or only directors who will never sit with a claims processor.
Selection rules that survive contact with a real company:
Coverage, not clones. At least one champion per first-wave team or site, including shifts and languages where the work actually happens.
Willingness to be coached on responsible use. A high-usage person who pastes secrets into prompts is a risk, not a champion.
Manager written agreement before the badge goes on the intranet.
Term length. A six-month tour with a refresh beats a lifetime appointment that burns out.
Diversity of functions. A champion network that is 90 percent IT will not move nurses, dispatchers, or store managers.
Early adopters and champions overlap but are not identical. An early adopter may be a heavy Dashboard user who never teaches. A champion must teach. Promote the teachers.
Time allocation is the real budget
If champion work is "whenever you can squeeze it," you do not have a program. You have a volunteer club that vanishes at quarter close. A transformation leader negotiates protected time with the champion's manager as part of first-wave launch, not as a favor.
A practical design for a knowledge-worker champion (adjust for frontline shifts):
Protected hours on the calendar (for example, a half-day equivalent spread across the week) for office hours, community posts, and the council feedback packet.
Office hours that teams can book, plus drop-in sessions after a new Copilot capability lands.
No on-call rotation for IT incidents.
Backfill or workload relief if the person's day job is already at capacity. Praise is not a resource.
A warehouse operations VP who wants a second-shift champion must free that person from a full pick-rate target during the coaching window, or the champion will rationally ignore Copilot. Put the hours in the labor plan.
Community of practice
Microsoft's playbook treats the user community and the champion network as related but not identical. The community is the broad peer space. Champions are the people you have asked to lead inside it.
Learn's suggested starter kit:
A Teams space (the playbook mentions a group for the AI council to discuss goals; champions need their own working channel too).
Viva Engage so employees can share tips, ask questions, and upvote answers.
Microsoft Forms to log what is working and what is not, so the council can improve training and use cases.
Microsoft's onboarding toolkit posts, when you use them, so you are not inventing every launch message from a blank page.
That is a community of practice : repeated sharing of prompts, failures, and role scenarios. Cadence matters more than a fancy platform. A 30-minute fortnightly champion huddle plus a living prompt gallery in SharePoint will beat a gorgeous site nobody opens.
Plug champions into Microsoft's public Microsoft 365 Champion Program resources if they want extra peer networking and adoption content. That is optional enrichment, not a substitute for your internal time and manager support.
Feedback loop to the council
Champions are the council's field sensors. Without a loop, the council debates slogans while the floor already knows Copilot is citing a retired price list.
Run a simple packet every council meeting:
Top three friction themes from office hours (permissions, missing files, confusing SKUs, a feature people think exists).
Top three winning prompts with the business outcome, ready for comms to celebrate.
Responsible-AI or data near-misses (a prompt that almost leaked; a hallucinated policy number someone almost sent to a customer).
Requests that belong to IT (license, DLP, agent controls) versus requests that belong to the business (process change).
Champion health : who is overloaded, which sites have no coverage, which managers revoked protected time.
The champions liaison owns the packet. Champions do not need to sit in every two-hour council meeting. Sending them all into the governing table turns coaches into politicians.
Measuring champion-program health
If you only measure Copilot active users, you cannot see whether the champion layer is working or whether IT got lucky with a curious department. Track a short scorecard:
Signal Healthy Unhealthy
Roster Named champions with manager agreements and end dates Anonymous "anyone can be a champion" with no owners
Time Protected hours actually taken (calendar proof) Coaching only after 6 p.m.
- Coaching Office hours held; attendance from the intended teams Empty sessions, or the same three friends
- Community Posts, replies, upvotes, prompt gallery growth A dead channel with a launch GIF
- Escalation quality Specific issues the council or IT can act on Vague "people are unhappy"
- Load Coaching volume the liaison can see; backup champions One hero doing 80 percent of sessions
- Turnover Planned rotations Quiet quits after the first incident
- Scope Prompts and workflows Ticket queues, license assignment, DLP exceptions
Dashboard usage still matters: Learn says you can use it to spot power users to recruit. It is an input to selection, not the definition of champion success. A quiet but trusted nurse who coaches a ward may not top the tenant-wide usage chart and can still be your best champion.
Scenario: the helpdesk dump
An IT director announces that all "How do I Copilot?" tickets will auto-route to departmental champions so the service desk can stay focused on outages. Within a month, champions are resetting MFA, explaining why a file is not in Copilot (oversharing), and missing their day jobs. Two resign the badge. The director created a shadow service desk. The fix: return access and outage tickets to IT; publish the champions-versus-IT table; restore protected coaching hours; have the liaison report champion load to the council. Copilot questions about prompts stay with champions. Tenant problems stay with IT.
Scenario: legal wants a champion who will police prompts
A general counsel asks for a "compliance champion" in every division who must approve prompts before staff may use Copilot. That design will freeze work and make champions hated. The leader alternative: council-owned data rules and DLP; champions who teach those rules with examples ("do not paste this contract clause into a consumer bot; here is the in-tenant pattern"); spot audits by legal or risk, not a champion bottleneck on every sentence. Champions amplify policy. They do not become a second legal department.
When selection, time, community, feedback, and measurement are in place, the champions program is established. When any one of those is missing, you have a slogan. AB-731 is testing whether you can tell the difference.