We deployed D365 Finance & Operations 2025 Wave 2 to production three weeks ago. After the dust settled, I wanted to share what actually matters - not the marketing slides, not the release notes, but the features that changed how we work versus the ones that are still waiting for their moment.
Context: we're a $110M food distributor running D365 F&O with heavy catch weight usage, integrated with BFC Dakota WMS, Adobe Commerce for B2B, and Shopify Plus for B2C. Our use case is specific but not unusual for the food distribution industry.
Features That Delivered Real Value
Copilot in Finance - Natural Language Queries
The ability to ask questions like "Show me all overdue invoices from customers in the restaurant segment" and get actual results is game-changing. Our AR team went from running reports to having conversations with the system. It's not perfect - complex queries sometimes misinterpret intent - but for 80% of ad hoc questions, it's faster than building a report.
Verdict: High value, daily use
Enhanced Catch Weight Processing
This was the update I'd been waiting for. The improvements to catch weight tolerance handling and the new validation rules for weight capture at receiving have reduced our weight discrepancy exceptions by 34%. That translates directly to fewer billing disputes and faster invoice processing.
Verdict: Critical for food distribution, major improvement
Improved Dual-Write Performance
Our integration with Dataverse and downstream Power Platform apps had been experiencing lag during peak hours. The Wave 2 improvements to dual-write sync performance were immediately noticeable - sync latency dropped from 8-12 seconds to under 3 seconds for most tables. Our real-time inventory visibility in Power Apps is now actually real-time.
Verdict: Significant operational improvement
Warehouse Management Mobile App Updates
The new UI for mobile warehouse operations is cleaner, faster, and requires fewer taps for common workflows. Our pickers noticed immediately. Training time for new warehouse staff has dropped because the interface is more intuitive. Small improvements that compound across thousands of daily transactions.
Verdict: Incremental but meaningful
Features That Fell Flat
AI-Powered Demand Forecasting
We had high hopes for this one. The reality: it needs significantly more historical data than we have, and when we fed it what we have, the forecasts were less accurate than our existing statistical models. Microsoft says it improves over time, but we've turned it off for now. Maybe revisit in 6 months.
Verdict: Not ready for our use case
New Financial Dimensions Experience
The redesigned financial dimensions interface is supposed to be more intuitive. Our finance team finds it slower and more confusing than the old version. We've submitted feedback through our partner, but for now we're using the classic view where possible.
Verdict: Step backward for experienced users
Embedded Power BI Refresh
The embedded Power BI experience got visual updates but still doesn't solve the fundamental issue: refresh frequency limitations for embedded reports. Our users still prefer going to the full Power BI service for anything that needs data fresher than hourly.
Verdict: Cosmetic improvement, core limitations remain
What We're Still Waiting For
Every Wave release comes with features that are "planned" or "coming soon." Here's what we actually need:
Better lot tracking for food safety. The current lot tracking works but requires too many manual steps for trace-forward and trace-back scenarios. FDA requirements keep getting stricter, and we need more automation here.
Native Adobe Commerce connector. We're running custom middleware for our B2B e-commerce integration. A native connector would reduce maintenance overhead and improve reliability. Microsoft has hinted at this but no timeline.
Improved catch weight support in Planning Optimization. Planning Optimization is the future, but its catch weight handling is still behind the deprecated Master Planning. We can't fully migrate until this gap closes.
The Update Process Itself
I've written before about Microsoft's relentless update cadence. Wave 2 deployment went smoothly for us, but only because we've built discipline around it:
- Sandbox testing: 3 weeks in sandbox before production, with documented test cases covering our customizations
- RSAT automation: 847 automated regression tests that run overnight after any update
- Partner review: Our implementation partner reviews release notes for anything affecting our custom code
- Rollback plan: Documented and tested, though we haven't needed it yet
Total internal effort for this update: approximately 120 person-hours across IT and business testing. That's significant, but it's predictable and we've budgeted for it.
Copilot: A Deeper Look
Since Copilot is the headline feature, let me go deeper on what's working and what isn't.
What works:
- Simple queries against single entities (customers, vendors, items, transactions)
- Aggregations and summaries ("total sales this month by customer segment")
- Status checks ("which POs are pending approval")
- Natural language filtering ("show me customers who haven't ordered in 30 days")
What doesn't work yet:
- Complex joins across multiple entities
- Anything involving our custom fields (it doesn't know about them)
- Comparative analysis ("compare this month to last year")
- Anything requiring business context it wasn't trained on
The key insight: Copilot is excellent for operational queries that would previously require report building. It's not a replacement for proper analytics or business intelligence. Manage expectations accordingly.
Recommendations for Other CIOs
If you're planning your Wave 2 deployment, here's my advice:
Don't rush Copilot adoption. It's genuinely useful, but it needs training time - both for the AI and for your users. Roll it out to power users first, gather feedback, then expand.
Test catch weight thoroughly if you use it. The improvements are real but behavior changes could affect your workflows. We found two edge cases that required process adjustments.
Plan for dual-write if you haven't already. The performance improvements make it viable for scenarios that were previously too latency-sensitive. If you've been avoiding Dataverse integration, this might be the wave to reconsider.
Budget for testing. Every Wave release needs proper testing. If you're not allocating time for this, you're accumulating risk. The question isn't if an update will break something, but when.
Looking Ahead to 2026
Microsoft has previewed some of what's coming in 2026 Wave 1. I'm particularly interested in the expanded AI capabilities in supply chain planning and the continued evolution of Copilot. But I've learned to evaluate features when they ship, not when they're announced.
For now, we're stable on Wave 2, our users are adapting to the new capabilities, and we're positioned well for whatever comes next.
Planning an ERP modernization?
6 SAP-to-Dynamics conversions with zero business disruption. Let's discuss your project.
ERP Services Book a Call