What’s Coming to Dynamics 365 Business Central in 2026

AI Agents: The Beginning of Intelligent ERP

The biggest shift coming in 2026 is the introduction and expansion of AI agents within Business Central.

Now, the word “agent” gets thrown around a lot. But in this context, it means something specific.

An agent inside Business Central is not a chatbot.
It’s not a macro.
It’s not a scripted workflow.

It’s an AI-powered assistant that can interpret high-level business goals written in natural language and translate them into structured, multi-step ERP actions.

That distinction matters.

Traditional ERP automation works like this:
“If X happens, then do Y.”

Agents work like this:
“Here’s the goal. Understand the context. Determine the steps. Execute within boundaries. Ask for approval where required.”

That’s a completely different paradigm.

These agents reason over Business Central data. They understand structured financial and operational records. They sequence dependent actions carefully — especially when financial consequences are involved. And critically, they operate strictly within assigned permissions.

Every action an agent takes is logged. Every decision can be reviewed. Sensitive operations require human approval.

This isn’t AI replacing control.

It’s AI operating under control.

Built-In Agents: Sales and Payables Reimagined

At launch, Business Central introduces built-in agents like the Sales Order Agent and the Payables Agent.

Think about how many businesses still manually process incoming sales instructions via email. Or how many finance teams still extract invoice details and match them against purchase orders line by line.

Those are high-volume, repetitive tasks that consume skilled employees’ time.

The Sales Order Agent can interpret natural language instructions or email content and transform them into structured sales documents. The Payables Agent can extract vendor invoice details and match them against purchase orders, dramatically reducing manual reconciliation.

But the bigger story isn’t the individual use cases.

It’s the runtime layer behind them.

Business Central is introducing an agent runtime framework embedded directly into the ERP core. This runtime ensures that agents respect company boundaries, environment boundaries, tenant isolation, and user permissions. It provides a timeline view of every step taken. It supports memory across multi-step execution. It enforces human-in-the-loop governance.

ERP systems are inherently stateful and financially consequential. Introducing AI into that environment requires discipline.

This runtime is that discipline.

Designing Your Own Agents: A New Innovation Surface

Beyond built-in agents, 2026 introduces the ability to design your own custom agents directly inside Business Central.

Through the AI Development Toolkit, consultants, developers, product owners, and even advanced users can begin prototyping intelligent agents tailored to their specific business processes.

This is where things get interesting.

Instead of writing complex AL code for every automation scenario, you define agent goals and instructions in natural language. You refine them. You test them in sandbox environments. You assign profiles and permissions. You inspect execution logs to understand reasoning paths.

Agent definitions can be exported as structured JSON files and version-controlled in systems like GitHub, enabling governance and change tracking.

This lowers the experimentation barrier dramatically.

It allows partners to envision new ERP-native AI use cases without immediately committing to full production deployments.

And while custom agents consume Copilot credits, the billing model is structured through prepaid or pay-as-you-go frameworks — introducing scalability without unpredictable complexity.

This is not just a feature.

It’s a platform capability.

Managed AI Resources: Simplifying the AI Stack for ISVs

Another major 2026 shift happens behind the scenes.

When publishers extend Copilot capabilities in Business Central, they can use Microsoft-managed Azure OpenAI resources by default.

Why does this matter?

Because AI infrastructure management is not trivial.

Without this model, customers would need to provision Azure OpenAI subscriptions, deploy models, manage scaling, monitor usage, and ensure compliance.

That creates friction. It slows adoption. It complicates support.

With Microsoft-managed AI resources, publishers can extend Copilot functionality without forcing customers into infrastructure decisions. Scaling, load balancing, throttling, and security protections are handled centrally.

There’s unified AI consumption tracking. There’s cohesive billing across AI-enabled solutions.

For most ISVs, this default configuration will be more than sufficient.

However, if a publisher requires fine-tuned models or specialized deployments not covered by the managed resource pool, they can still bring their own Azure OpenAI subscription.

The flexibility remains.

But the complexity becomes optional.

That’s a meaningful strategic shift.

Excise Tax Automation: Quietly Powerful Compliance Infrastructure

Not every 2026 update is AI-focused. Some are foundational.

One of the more understated but impactful updates is the new excise tax calculation framework.

Excise taxes vary by product type, region, and regulatory requirements. They often change. They can carry significant penalties if miscalculated.

Manual handling introduces risk.

The new framework automates excise tax calculation logic and provides configurable support for different models. While it doesn’t initially include specific compliance reporting templates, it establishes the engine required to handle diverse regulatory frameworks.

For manufacturers and distributors operating in regulated industries, this reduces exposure and increases reliability.

Compliance automation isn’t flashy.

But it’s strategic.

Poland Data Residency: Governance as a Platform Feature

Another 2026 update reflects the growing importance of cloud governance.

Business Central environments using Polish localisation will move to the Poland Azure Geo to meet local data residency requirements.

For businesses operating in Poland, this ensures regulatory alignment and improved performance through reduced latency.

Importantly, the migration is handled by Microsoft. There’s no functionality loss. No operational disruption.

It’s governance embedded into the platform rather than bolted on afterward.

That signals maturity.

Quality Management: Strengthening the Supply Chain Core

Supply chain resilience is no longer optional.

The new Quality Management extension allows businesses to embed structured quality checks directly into receiving, production, and assembly workflows.

Organizations can define quality parameters such as physical measurements, sampling models, quarantine procedures, and certificate requirements. Quality checks can trigger automatically upon posting receipts or production outputs, or they can run on scheduled intervals.

This brings structured traceability and compliance into operational processes without external add-ons.

And because the extension installs automatically in new environments, adoption friction is minimized.

Quality becomes a native capability, not a patchwork solution.

The Strategic Pattern Behind 2026

If you step back from the individual features, a clear pattern emerges.

Microsoft is strengthening three pillars simultaneously:

Intelligence.
Compliance.
Governance.

AI agents introduce reasoning into ERP workflows.
Managed AI resources simplify AI adoption at scale.
Excise tax automation reduces regulatory exposure.
Data residency adjustments reinforce trust.
Quality management enhances operational reliability.

This isn’t feature expansion.

It’s architectural reinforcement.

Final Perspective

2026 will likely be remembered as the year Business Central transitioned from a transactional system to an intelligent execution platform.

Not because it added AI.

But because it embedded AI within controlled, auditable, permission-aware ERP structures.

The difference is subtle but profound.

ERP is no longer just recording what happened.

It’s beginning to participate — responsibly — in making things happen.

And that changes the trajectory of how mid-market businesses operate moving forward.