[object Object]

A platform owner reads the wave announcement, screenshots three slides, and emails the team. By Friday, three different stakeholders have asked when each feature lands and what it costs them to enable. The wave is dense; the operator’s read separates the announcements that need work this quarter from the ones that need work in two.

Release Window

Wave 1 rolls out April through September 2026. Announced March 18, 2026. Feature deployment continues in stages; consult the individual feature’s target date before planning dependencies. The release planner page is the source of truth for per-feature dates; the keynote slides are marketing.

April 2026: early-deploy features GA
June 2026: mid-wave landing batch
September 2026: end-of-wave, remaining features GA

Subscribe to the M365 message center filter for D365 so the per-feature dates land in your inbox automatically.

Sales

Sales Agent becomes the daily command center for sellers — conversational data access, mobile and in-context experiences, deeper configuration and governance. CRM data plus Microsoft 365 signals (email, meetings) power insights. The biggest configurable change is the per-role agent capability matrix, which lets admins tune what the agent can read and write per sales role.

agent_capabilities:
  Account Executive:
    read: account, opportunity, contact, activity
    write: opportunity_notes
  SDR:
    read: lead, contact
    write: lead_qualify

Map your sales roles before enabling, otherwise every rep gets every capability.

Customer Service and Contact Center

Agentic capabilities across case management, email, customer intent detection, quality evaluation, knowledge. Contact Center gets AI-powered self-service, accelerated assisted service, and operational intelligence for managers. The deflection metric finally gets a usable definition that excludes hangups and abandoned chats.

The auto-close threshold is the most contested setting. Default 0.92 is conservative; below 0.85 starts to leak misroutes to customers. Tune in 0.01 increments based on weekly misroute review.

Business Central

Accelerates toward agentic ERP — AI agents automating sales and purchase scenarios, alongside developer productivity investments for extensibility. Small-to-mid org ERP becomes meaningfully AI-first. The order-from-email flow handles the messy attachments that older parsers stumbled on.

purchase_policy:
  auto_approve_below: 500
  require_approval_above: 5000
  require_dual_approval_above: 25000

Define the policy explicitly. The agent will not invent thresholds for you.

What to Pilot vs Defer

Use a two-by-two: business value high or low, deployment risk high or low. Pilot the high-value low-risk features in May. Watch the high-value high-risk features through the mid-wave bug pass. Skip the low-value features entirely.

Pilot now: case triage, knowledge drafting, sales agent mobile
Watch: agentic email response, BC purchase orchestration
Skip: cosmetic UI changes, niche industry add-ons

Cross-Tenant Coordination

If you operate multiple tenants, coordinate the pilot tenant ahead of production by 30 days. Breaking changes hit the pilot first and you have a full sprint to react before they land in revenue-bearing environments.

Measuring the Wave

Define one or two metrics per pilot feature before you enable. For sales agent, measure activity logging volume and response time. For service triage, measure misroute rate and time-to-first-touch. Without baseline metrics, the post-wave review is opinions, not evidence.

What to do this week

Build the two-by-two for your modules, subscribe to the message center filter, and book the pilot environment for the May features. Define the metrics for each pilot before the feature lands so the comparison data is clean.

[object Object]
Share