Inside Microsoft's Employee Comms Playbook | Microsoft 365 Community Conference

Overview: What Microsoft is doing with an “employee comms playbook”

• Microsoft describes its employee/executive communications function as highly decentralized: a small corporate center (the “center” is limited) with many leaders and communicators distributed across countries and organizations.
• As a result, the “playbook” is focused less on centralized control and more on community enablement, upskilling, and repeatable ways to solve communication problems.
• AI adoption is presented as a human-in-the-loop approach:
  • AI accelerates analysis and drafting
  • People own prompts, questions, data quality, and the final decisions

Organization & Operating Model (Decentralized Community)

• Geek at Microsoft (GEK) is Microsoft’s internal community acronym for global executive & employee communications.
• The corporate center is small—cited as “10 people or so”—with communicators distributed globally.
• A key strategic objective is to share learnings and priorities across the community so best practices scale beyond the corporate center.

Upskilling and Community Education (Scaling Capability)

Scenario-based programs

The community participates in scenario-based training, including:

• Events program execution: latest tools/technologies/strategies for running events
• AI upskilling + Copilot/agent usage, including:
  • Using Copilot
  • Using agents
  • Building custom agents to scale work “from the center into the business”
• Education is modeled as scenario-based experimentation, where comm teams test approaches in realistic contexts.

Scale

• Upskilling is cited at “about a thousand people” in the broader effort (not all directly reporting to the speakers, due to decentralization).

AI + Copilot Intersection in Comms Workflows

Mindset: start with the workflow, not the tool

• The starting point is workflows/processes and the problem to solve, not “which tool should we use?”

Where AI is used

AI is applied at multiple levels:

• Individual daily work habits (speed and quality for comm tasks)
• Process/workflow improvements across the business
• Scaling key messages and narrative alignment
• Faster insights from employee/communications data

Examples of AI-enabled activities

Post-event analysis for CEO town halls

• Multiple approaches were tested, including:
  • Copilot-based analysis
  • A community agent
  • A PR agency approach
  • Survey/form-survey reporting
• Learning: while outputs shared common themes, they were noticeably different across methods—reinforcing the need for careful:
  • prompting
  • question design
  • data inputs

Employee listening / social signal analysis

• Shift from manual “bean counting” (likes/engagements) to automated, deeper AI analysis
• Goal: understand how things are landing to inform:
  • what to say
  • where to land it

Channel Strategy: How Microsoft Decides “What Goes Where”

Core principle

• First determine what problem is being solved, then decide:
  • what data/insights to use
  • which voices (e.g., leader voices)
  • whether the goal is broadcast or conversation

Decision flow (implicit in the talk)

• Problem definition → insights on employee needs + desired business change → message voice selection → channel choice

Moving from “where to post” to “what interaction you want”

• The emphasis shifts from:
  • “Should this be in email/SharePoint/Viva Engage?”
• To:
  • “Do you want conversation on this topic?”
  • “Do you want feedback and public Q&A?”

Two-way dialogue strategy

• A major trust-building lever is two-way dialogue
• Leader messaging routed into Viva Engage to:
  • drive transparent conversation
  • generate sentiment reporting back to inform next steps

Leader Engagement and Risk Management (Lessons Learned)

Reality: leaders differ

Leaders vary in:

• authentic communication style
• adoption rates for new tools/technologies
• comfort level and risk appetite

Comms-side competencies required

• Provide coaching so leaders can handle dialogue effectively
• Comms professionals act as trusted advisors, particularly for challenging topics (examples):
  • War in the Middle East
  • Company layoffs
  • Compensation topics
• Balance the need for:
  • transparency
  • effective question handling
  • legal risk protection
  • reputation management

Actionable Recommendations (What “Good” Looks Like)

Do

• Trust but verify (human judgment remains the owner)
• Use AI as:
  • a thought partner
  • a creative brainstormer
  • a method/testing assistant
  • a “poke holes” reviewer of your argument
• Experiment quickly—value becomes obvious once teams use it
• Identify areas with manual labor and apply AI to save time and free people for higher-value work

Don’t

• Never just copy and paste from Copilot
• Avoid drafting and immediately pasting without review

Metrics / KPIs Mentioned

• No explicit quantitative business KPIs were stated (e.g., revenue, CAC, LTV, churn).
• Communications evaluation metrics were referenced qualitatively, such as:
  • Post-event: sentiment and “how did it land”
  • Listening: earlier manual tracking of likes/engagements (“bean counting” era)
  • Current: automated/deeper analysis of employee response to guide decisions

Presenters / Sources

• Allison Michaels — Viva Engage product group
• Amy Morris — Microsoft global employee and executive communications team
• John Cerrone — Microsoft “Geek at Microsoft” (executive/employee comms community)