Updated · Originally published
Artificial intelligence is no longer a buzzword in ERP—it’s becoming a developer’s daily companion. With Microsoft embedding Copilot across Dynamics 365 Business Central, developers and partners are witnessing a shift from manual configuration to intelligent automation, predictive coding, and contextual insights.
For organizations exploring smarter ERP implementations, dynamics business central consulting team helps you leverage Copilot and AI innovations while keeping extensions upgrade-safe and future-ready.
This post dives deep into how AI is changing Business Central development, what’s new in 2025, and how partners can design, secure, and extend these features responsibly.
Copilot in Business Central: Capabilities and Developer Impact
Microsoft’s Copilot for integrates generative AI directly into daily workflows. End users can generate item descriptions, summarize records, and automate routine data entry. But for developers, it opens a new layer of extensibility.
Copilot lives inside the Business Central ERP environment—accessible through pages, AL code, and the upcoming Copilot Studio. Developers can define what data Copilot accesses, how prompts are structured, and where AI suggestions appear. Unlike generic AI assistants, this one is business-context aware, respecting company permissions and localization.
What’s New in 2025 (Wave 2): Why It Matters
The 2025 release introduces AI extensibility patterns inside AL, expanding the System.AI module. It also tightens Copilot integration with the Power Platform—allowing natural-language automation via Power Automate.
Meanwhile, Microsoft’s broader Copilot updates (like multi-model “smart mode” and context length improvements) mean richer ERP-specific prompts and better reasoning for business data. Developers can expect more accurate and domain-aware responses when building Copilot capabilities.
???? Pro Tip: Always test new features in a sandbox linked to telemetry so you can measure AI accuracy and cost before enabling it in production.
How AI Changes the Business Central Development Lifecycle
Traditional Business Central customization involved heavy AL coding, manual testing, and tedious debugging. AI introduces new efficiencies across that lifecycle.
1. AL Coding with GitHub Copilot
GitHub Copilot now supports AL language patterns. Developers can type natural prompts like “create an event subscriber for sales order release” and Copilot generates boilerplate code automatically.
Example benefits:
- Consistent syntax and naming conventions
- Fewer manual typos and runtime errors
- Time savings in repetitive object definitions
However, human review remains essential—AI suggestions should always pass peer review, test cases, and static code analysis before deployment.
2. Building Smarter Extensions
Copilot enables developers to create AI-driven extensions that respond intelligently to context. For example, a warehouse management app could include a Copilot that explains inventory anomalies or generates restock recommendations.
Developers must think in prompt-driven design terms:
- Define input (prompt) and output (response) clearly
- Use structured output (JSON or AL records)
- Maintain idempotent logic—AI results shouldn’t alter core data unpredictably
3. Testing, Debugging, and Telemetry
AI doesn’t remove the need for testing—it redefines it.
Use AL test libraries to validate AI outputs and Application Insights telemetry to monitor prompt success, response quality, and latency.
Telemetry can reveal:
- Prompt length vs. response accuracy
- Cost implications of token usage
- Unusual AI behaviors in production environments
This data helps you refine both AI logic and performance.
How to Extend Copilot in Business Central (Developer Walkthrough)
Business Central’s System.AI module allows partners to register custom Copilot capabilities through AL. Below is a simplified conceptual snippet to illustrate:
codeunit 50100 "Folio3.AI.CopilotExample" { [System.AI.Capability('SuggestInvoiceNote', DisplayName = 'Generate Invoice Note', Description = 'Creates a personalized note for overdue invoices')] procedure SuggestInvoiceNote(CustomerName: Text; OverdueDays: Integer): Text var prompt: Text; response: Text; begin prompt := StrSubstNo('Write a polite reminder for %1, overdue by %2 days.', CustomerName, OverdueDays); response := System.AI.Run(prompt); exit(response); end; }
This registers a custom Copilot capability that generates dynamic invoice notes based on input variables.
In production, you’d secure this with data permissions, structured outputs, and telemetry hooks.
For a deeper dive into implementation guidance, explore our , where we showcase how AI can automate and optimize Business Central workflows.
Security, Governance, and Performance
AI development brings new risks—data exposure, unreliable outputs, and compliance gaps. Here’s how to manage them effectively.
|
Risk / Concern |
Example in Business Central |
Recommended Mitigation |
|
Data leakage |
Copilot pulls data across companies | Apply strict permission sets and environment isolation |
| Unreliable responses | Hallucinated business text |
Use structured outputs + human approval workflow |
|
Lack of audit trail |
No logs of AI outputs | Enable prompt logging in Application Insights |
| Latency or token cost | Long prompts / retries |
Optimize prompt templates and monitor token usage |
Best Practices:
- Always validate Copilot responses before saving to production tables.
- Keep AI features opt-in for each environment.
- Document each capability’s business purpose and data access scope.
High-Impact AI Use Cases for Business Central Partners
You don’t have to build a massive AI engine to deliver value.
Start with narrow, high-ROI Copilot scenarios:
- Collections Management: Automatically draft polite overdue reminders using customer tone preferences.
- Procurement: Generate item descriptions from supplier data or specs.
- Financial Analysis: Summarize long account schedules into human-readable insights.
- Inventory Planning: Explain sudden stock variances and suggest reorder actions.
These examples combine AL extensibility with AI assistance—providing measurable productivity gains while keeping human control.
Choosing the Right Dynamics 365 Business Central Partner
Building Copilot features requires both technical and governance expertise. When evaluating a Dynamics 365 Business Central partner, ensure they:
- Have hands-on experience with AL + System.AI module
- Understand data compliance and telemetry in Azure
- Can design prompt templates aligned with your business domain
- Provide a structured sandbox testing and deployment process
At Folio3, we combine technical depth with domain knowledge to help organizations pilot and scale Business Central implementation enhanced by AI.
Conclusion: AI Won’t Replace BC Developers—It Will Amplify Them
AI and Copilot are redefining what it means to develop for Microsoft Dynamics 365 Business Central.
From AL coding to user experience, automation to analytics, AI acts as a force multiplier—empowering developers and consultants to deliver faster, smarter, and safer ERP solutions.
As Copilot evolves, those who adapt early will shape the next generation of intelligent ERP.
If you’re ready to explore practical Copilot scenarios tailored to your business, connect with our Dynamics 365 Business Central consulting team to see what’s possible.