As organizations move from AI curiosity to real implementation decisions, the conversation inevitably shifts toward data. Not just whether it’s clean, but whether there’s a clear Business Central data strategy behind it.
In the first two posts of this series, I focused on readiness — what Microsoft Copilot can help with and how to assess whether your organization is prepared. But once that groundwork is in place, deeper questions emerge:
How is your ERP data structured?
How is it shared?
And how does it support more advanced AI scenarios across the Microsoft platform?
Copilot works within the boundaries of your data environment. The quality and structure of that environment ultimately determine how far AI can go, and that’s exactly where a well-defined Business Central data strategy begins to matter.
This post explores what that means in practical terms.
Does Copilot work if my data isn’t clean?
Technically, yes. Copilot will still generate summaries, explanations, and suggestions. But the value becomes inconsistent.
Copilot draws on the transactional and master data inside Business Central. If that data is duplicated, inconsistently defined, or fragmented across departments, AI outputs will reflect those realities. It won’t correct them.
This is where many organizations confuse functionality with effectiveness. The feature works. But the outcomes feel uneven because the underlying Business Central data quality isn’t consistent enough to support reliable insight.
That’s why ERP data readiness for AI isn’t just about enabling Copilot. It’s about ensuring the environment Copilot operates in is structured intentionally… something that should be guided by a clear Business Central data strategy rather than reactive cleanup efforts.
What data does Copilot use in Business Central?
Copilot uses the data available within your Business Central tenant: financial transactions, master records, dimensions, permissions, and historical context. It operates within existing security boundaries and respects role-based access.
But Copilot only sees what’s connected and structured properly.
If Business Central functions as an isolated ERP database, Copilot’s insight is limited to what lives inside that system. If Business Central is part of an integrated ERP data platform with shared definitions and consistent governance, the quality of those insights improves dramatically.
That difference isn’t accidental. It’s the direct result of whether your Business Central data strategy was designed to support scale beyond day-to-day transactions.
Why integrated systems aren’t the same as shared data
Many organizations describe their environment as “integrated.” Systems pass information back and forth. APIs connect applications. Reports pull from multiple sources.
But integration is not the same as a shared data foundation.
Integrated systems can still maintain conflicting definitions, duplicate records, or inconsistent dimensional structures. A shared foundation, by contrast, aligns data models and governance so that systems operate from the same source of truth.
This difference matters for AI.
Research from McKinsey highlights that organizations struggle to scale AI when ERP systems and data models aren’t aligned to support it enterprise-wide. Without a unified data foundation, AI initiatives remain siloed and difficult to extend across the organization.
When leaders pursue AI without addressing siloed data ERP challenges, they often discover that “connected” systems don’t necessarily produce coherent insight. That’s why a disciplined Business Central data strategy must prioritize shared definitions over surface-level integration.
Why your Business Central data strategy determines AI success
A thoughtful Business Central data strategy defines more than data cleanliness. It defines ownership, modeling standards, governance policies, and how ERP data participates in the broader Microsoft ecosystem.
This is where the concept of an AI data foundation ERP becomes practical. It means:
- Consistent master data definitions
- Clear dimensional alignment
- Reduced duplication
- Governed data flows across systems
- Intentional integration with analytics platforms
Without this strategy, AI remains constrained to surface-level assistance. With it, organizations unlock the ability to expand into forecasting, predictive analytics, and cross-functional intelligence.
A recent Forbes article reinforces this shift, noting that organizations preparing ERP systems for AI must focus on data structure and interoperability, not just feature adoption.
AI outcomes are rarely limited by features. They’re limited by foundation — and foundation is exactly what a strong Business Central data strategy is designed to address.
How does Business Central fit into Microsoft’s data platform?
Business Central doesn’t operate in isolation. It sits within a broader Microsoft ecosystem that includes Microsoft 365, Dataverse, and increasingly, Microsoft Fabric.
When Business Central data connects intentionally to these platforms, organizations gain access to deeper analytics and advanced AI scenarios.
For example:
- Dataverse Business Central integration enables shared data models across applications.
- Microsoft Fabric Business Central analytics supports large-scale data processing and enterprise-level insight.
- Unified governance ensures consistency across reporting, forecasting, and operational analysis.
Microsoft’s documentation on connecting Dataverse environments to Fabric illustrates how shared data foundations enable broader analytical capabilities across the platform:
These capabilities deliver value most effectively when they are aligned to a cohesive Business Central data strategy, rather than layered on top of fragmented data environments.
What data issues block AI value the most?
Certain patterns consistently limit AI effectiveness:
- Duplicate customer or vendor records
- Misaligned dimensions across departments
- Shadow reporting layers outside Business Central
- Conflicting KPI definitions
- Inconsistent data ownership
Individually, these issues may seem manageable. Collectively, they undermine trust.
When trust declines, adoption slows. When adoption slows, ROI follows.
That’s why ERP data readiness for AI and long-term architecture planning must move together. Fixing isolated data problems is helpful. Establishing a durable Business Central data strategy is transformational.
From readiness to measurable outcomes
This series has intentionally moved in phases.
The first post established the strategic context for AI in Business Central.
Then we focused on assessing readiness.
This final post explains the data foundation required to make AI effective at scale.
Together, these pieces reinforce a simple idea: AI value isn’t unlocked by enabling features alone. It’s created through progression — strategy, readiness, and architecture working together.
And remember: AI in Business Central isn’t a single switch. It’s a structured journey. And like any journey, the direction matters as much as the destination.
About the Author
Adam Drewes is the Chief Technology Officer at Kopis, where he helps companies make smarter software decisions that align with their business goals, whether that means deploying proven tools or building custom solutions that protect their competitive edge.
With more than two decades in the software services space, Adam brings a rare mix of technical depth and business insight to every conversation. He’s endlessly curious about how companies operate, what drives their success, and how the right technology choices can accelerate their growth.
