For decades, business applications have been used by organizations as systems of record capturing what already happened. Opportunities were updated after the call. Orders logged after the decision. Journal entries posted after the close. While that model was reliable, it was inherently reactive.
This is now changing, as intelligence is built into business applications they are becoming systems of action, where AI agents don’t just record work, they help carry it out, with people setting the guardrails. For organizations using Dynamics 365, this is a shift from systems that primarily record business activity to systems that help move work forward under human-defined controls.
Making this practical takes more than agent reasoning. Agents need a governed way to access business context, connect to transactional systems, and act through the same rules, permissions, and audit trails already built into the application. That is the role Dynamics 365 Model Context Protocol (MCP) servers play.
What Dynamics 365 MCP servers change for your business
Dynamics 365 MCP servers make the business applications platform agent-ready. Agent-ready means agents can connect to and operate inside the transaction layer of the business application, using the same data models, rules, permissions, security guardrails and audit trails as any user of the system.
Instead of integrating AI around enterprise resource planning (ERP), organizations can bring agents into the flow of work, where business decisions and transactions already happen:
- Faster cycle times, by moving approvals, forecasts, and transactions forward with Dynamics 365.
- Lower integration overhead, by using the application layer customers already govern instead of building custom connections for every process.
- Reduced manual effort, as agents clear routine steps while people stay in control of thresholds and approvals.
- Better-informed planning, as agents work from trusted operational data and planning context.
Today, Dynamics 365 exposes over 650,000 MCP actions across sales, finance, supply chain, human resources, field service, customer service, and project operations. That is the same depth of capability users rely on inside the application, now available to agents.
When an agent qualifies a lead, adjusts a forecast, or processes a transaction, it runs within the Dynamics 365 transaction layer, using the same data, rules, permissions, and audit trails already in place. That reduces the need to build separate APIs or bolt-on governance layers for every scenario.
And given Dynamics 365 is MCP compliant, it can connect to compatible agents across supported platforms, enabling multi-agent interoperability, and customer choice, not lock-in.
How partners use MCP to build AI agents with Dynamics 365
Partners are at the front edge of this shift. They have industry expertise, domain knowledge, and customer relationships to turn platform capability into real-world outcomes.
Cegeka: AI in high-stakes, regulated workflows
In manufacturing and life sciences, Cegeka operates in environments where precision is not optional, it is regulated. Their Quality Impact Recall Agent is designed for manufacturing focused organizations managing quality incidents that propagate in two directions: backward through batch genealogy to determine what was produced and forward through customer commitments to identify who received it. When performed manually, this process can take hours or even days, according to Cegeka.
How do you introduce AI into a business process where 99% accuracy is considered a failure? In the rigid, transactional world of ERP, processes like product recalls leave no room for hallucinations or best guesses.
Thomas Van Dorpe, AI ERP Architect at Cegeka
Cegeka’s approach anchors the agent in ERP data, with a Traceability Add-in that maps batch and lot relationships across the supply chain. The agent can trace from a defect signal back to affected batches and forward to affected customers. Microsoft Copilot Studio orchestrates the agent logic, and MCP connects Dynamics 365 to Microsoft 365, where supplier communications, corrective workflows, and regulatory documentation get initiated.
Across Cegeka’s broader portfolio, their Tender Agent built on Microsoft Power Apps, Copilot Studio, and Dynamics 365 Project Operations helps procurement teams collect and compare supplier quotes, manage request for proposal (RFP) processes, and convert approved tenders into structured work. A Configuration Comparison Agent monitors compliance-relevant drift, while a Regression Testing Agent automates test creation for periodic release validation. The use cases differ; the architectural approach does not.
proMX: Scaling agents across project-based operations
A different pattern emerges with proMX, a partner with deep roots in Dynamics 365 Project Operations. Rather than building a single, monolithic agent, they are assembling a portfolio of capabilities aligned to how project-based organizations operate, including change orders, approvals, time and expense, project management office (PMO) support, and lessons-learned retrieval. Their focus is concrete: energy, utilities, construction, and consulting firms that manage delivery projects, track time at granular levels, and execute financial adjustments on a scale.
proMX is also using its own environment as the proving ground, serving as a customer zero for the Dynamics 365 Project Operations Activity Approvals Agent. They started with time-entry approvals where high volumes of short-duration entries were stalling billing prep and slowing month-end close.
The agent went from preview to full internal rollout in a matter of weeks, reaching 50% adoption across proMX’s internal PMO. Approvers stay in the loop, rules and thresholds are defined up front, and approvers see what the agent recommends before anything posts, while the repetitive work of clearing routine entries gets handled automatically. proMX reports that approval cycles shortened, billing prep accelerated, and month-end close becomes smoother.
For us, this was never about building an AI demo. Reaching 50% internal adoption shows that the Approval Agent is not just technically possible, it brings enough benefits for people to try and stick with it.
Sebastian Sieber, Global Director Technology, proMX
Their flagship in-market offering is the Change Order Agent. It handles a class of cross-record adjustments that typically break lighter-weight automation: reallocating budgets across projects, splitting labor and non-labor components, and preserving alignment between cost and sales. The agent interprets the request, maintains the integrity of underlying values, generates the required change-order records, and triggers the downstream system processes.
The architecture extends the Dynamics 365 stack. Copilot Studio orchestrates the agent logic, Microsoft Dataverse and Power Apps handle execution. And MCP connects the pieces, with proMX building targeted MCP servers only where gaps exist in the standard surface. For more complex, multiagent scenarios, Microsoft Foundry provides the scaling path. It’s a win on both sides: their unique value stays protected, and their solution keeps working as Dynamics 365 evolves.
TrueGradient: Expanding the ecosystem with AI-native companies
TrueGradient is an AI native company offering an integrated business planning platform spanning demand forecasting, inventory optimization, and price and promotion planning. It represents a growing wave of AI native companies building on Dynamics 365 and MCP, forming a new cohort within an evolving partner ecosystem.
Forecast Enrichment Agent handles the analytical heavy lifting, selecting an appropriate forecasting model for a given scenario, while a companion agent turns those forecasts into planning decisions. MCP serves as the connective layer into Dynamics 365 and the other ERP, data, and commerce systems customers already run.
Anchor customers such as Angelcare, Anand Sweets and more are live on Microsoft ERP and expanding in contract value. At Angelcare, a consumer packaged goods (CPG) manufacturer with baby and pet care brands sold in more than 50 countries, Angelcare reports a 12% improvement in demand forecast accuracy, with engagement targets that include 10–20% gains in trade promotion efficiency and roughly 10% reductions in working capital and logistics costs.
Our vision is to integrate AI-driven tools into our operations, and we selected TrueGradient for its cutting-edge Self-Serve AI platform and deep expertise in CPG and retail. We are seeing promising results.
Stephane Huot, VP Supply Chain & Logistics, Angelcare Group
The broader signal matters as much as the metrics. Companies built around AI, like TrueGradient, are choosing Dynamics 365 and MCP as the trusted place where their insights turn into real business actions at scale. That’s a sign the ecosystem isn’t just longtime partners updating what they offer, it’s growing in new directions.
Bring AI agents into the workflows your business already runs
The agent model is not theoretical, it is beginning to deliver impact across business processes and industries within our partner ecosystem. We are seeing agents move from experimentation to execution, with early results in real-world scenarios.
This is what agent-ready looks like in practice: agents connected to business data, governed by business rules, and trusted to take on real work. Not just assisting but acting in the system where work already happens.
Customers can gain faster time to value while operating within existing Dynamics 365 controls, the same permissions, security guardrails, and audit trails as any user. Partners benefit from a shorter path from idea to repeatable, AI-driven outcomes. And across the ecosystem, MCP serves as the connective tissue that makes agents interoperable, auditable, and production‑ready.
Start with the process that matters most
Explore partner-built agents already running on Dynamics 365 or identify the first manual process your organization can convert into an agent-driven workflow using Copilot Studio, Microsoft Foundry, and Microsoft Marketplace.
