Another wave of updates is coming to Dynamics 365. During the Microsoft Business Applications Update, it became clear that changes are planned across nearly all modules - from sales and customer service to supply chain, finance, HR, and field service. The common denominator: AI agents that autonomously execute operational work, escalating only exceptions and decisions that require human judgment. Below, I’ve outlined the most important updates.
Sales
The Sales Development Agent automates lead qualification and early-stage engagement. When a prospect attends a webinar or another event, the agent analyzes their role and company context, prepares a personalized message, and conducts a qualification conversation. Once the lead is ready for the next step, the agent hands it over to a sales representative with full context: conversation history, identified challenges, and recommended next actions.
Microsoft demonstrated this scenario using the example of a global professional services company with nearly 400,000 employees across 700 offices in 150 countries.
The Sales Close Agent supports sales teams in active customer relationships. In the demo, a client had just been acquired, bringing new leadership and shifting priorities. Within seconds, the agent builds a full picture: what changed post-acquisition, what external signals indicate new priorities, and what insights come from previous meetings and internal notes. It then generates a suggested agenda for the upcoming meeting. It also works as a Teams assistant - sales reps can ask questions in chat and receive answers grounded in the full relationship history.
A natural language pipeline view allows sales managers to ask about pipeline health, conversion distribution, or cross-sell potential without commissioning reports. The system responds based on agent activity data, handoff quality, and campaign-to-pipeline attribution.
Customer Service
Dynamics 365 Customer Service introduces a proactive model - the system initiates contact before the customer reaches out.
In a scenario presented by Zurich Insurance Hong Kong, a customer who recently extended her policy receives a proactive message explaining the new coverage. She replies with a question about renovation coverage - the agent verifies the policy and responds clearly. When she asks about limits and extension options, the conversation is handed over to a consultant. The consultant immediately sees the full context: customer profile, policy history, the proactive message, and the entire conversation.
The Quality Evaluation Agent automatically reviews all closed cases based on supervisor-defined criteria. For each case, it generates a score, explanation, and coaching recommendations. Supervisors can access results directly within the case view, without manually reviewing calls.
Proactive operational monitoring alerts supervisors with root-cause analysis before KPIs exceed thresholds. In the demo, an increase in average handling time was traced to incorrect routing - renewal-related cases were sent to general queues instead of specialists. After correcting routing, the issue was resolved.
Zurich Insurance Hong Kong had previously implemented Dynamics 365 Contact Center and Customer Insights, reducing email and call volumes. The new AI agent capabilities build on that foundation.
Supply Chain & Warehouse
The scenario presented involved a global aftermarket automotive company with multiple warehouses, hundreds of thousands of SKUs, and seasonal demand peaks.
Demand Planning, with promotional data integration, allows teams to upload campaign details - duration, SKU scope, discount level - directly into the forecasting model. Planners and pricing managers can assess the impact on sales volume, revenue, and supply chain before the promotion goes live.
Inventory positioning is automated - the system generates transfer orders based on demand forecasts. In the demo, high-turnover spring SKUs are moved closer to picking zones ahead of peak season. This is complemented by optimized picking routes that group and sequence orders to minimize distance and reduce worker congestion.
Finance
The Unified E-Invoicing Integration Framework enables integration with certified e-invoicing providers through configuration, without coding. The system automatically generates and routes e-invoices in compliance with local regulations, including support for prepayments and post-delivery settlements.
Note for Polish organizations: Microsoft did not explicitly confirm support for KSeF in this presentation. The framework is designed to integrate with certified local providers.
Organisations operating in Poland that wish to integrate their Dynamics 365 system with KSeF can make use of the ready-made ANEGIS solution. For more information, please visit the ANEGIS KSeF module page.
AI in accounts payable processing automatically recognizes supplier invoices, classifies header and line items, and matches them to purchase orders. Exceptions are displayed side-by-side for review. The system learns from user corrections, reducing future discrepancies.
The Account Reconciliation Agent automatically identifies transactions on inventory accounts and matches them with financial records. Only true discrepancies are escalated for human review.
B2B Commerce
B2B Multi-Outlet Commerce allows buyers to log in once and manage orders for multiple locations or organizations. Pricing, catalogs, and credit limits adjust automatically based on the selected account.
In B2B call centers, an AI agent integrated with Teams identifies the buyer, selects the correct account, and processes orders without switching systems. The agent operates on MCP (Model Context Protocol) - the same integration layer used by the self-service portal, ensuring a consistent experience across channels.
HR, Project Operations & Field Service
LinkedIn Recruiting enables job postings directly from Dynamics 365 Human Resources. A new role-based access portal replaces manual CV transfers between systems.
Multi-resource booking allows resource managers to assign several specialists simultaneously via the schedule board and confirm all bookings in a single action.
The Onboarding Agent automates document collection, training assignments, and onboarding checklists tailored to each role.
The Scheduling Operations Agent in Field Service enables real-time route optimization for up to five technicians or ad-hoc optimization for up to thirty resources. The mobile app includes a simplified timeline-based note field for adding text, photos, and videos without switching screens.
What this says about Microsoft’s direction
A clear pattern emerges: Microsoft is shifting from “systems of record” to “systems of action.” AI agents in Dynamics 365 operate autonomously within defined processes, escalating only exceptions and decisions requiring human judgment.
This trend requires a degree of caution. Agent autonomy depends on data quality. Organizations with fragmented CRM data, incomplete interaction histories, or weak master data management will not see the same benefits. Poor data can lead to confident but incorrect recommendations.
Microsoft is also positioning MCP as a core integration layer, enabling consistency across channels without separate integrations for each touchpoint.
FAQ
Can the Sales Development Agent replace an inside sales team?
No. It handles qualification and early nurturing. Negotiation and relationship-building remain human-led.
How does the Quality Evaluation Agent handle varied conversation contexts?
It follows supervisor-defined criteria, not generic standards. Edge cases are escalated for manual review.
Does integrating promotions into Demand Planning require custom setup?
No, it’s part of standard functionality. The level of accuracy depends on data quality.
When will these features be available?
They will roll out over the next six months in successive release waves.
Does the E-Invoicing Framework support KSeF?
Not explicitly confirmed. Integration via certified providers is possible.
