Microsoft Dynamics GP served a lot of businesses well. It was reliable, and for the better part of two decades it did exactly what a finance and operations system needed to do.
That period has ended, and AI has played a major role in changing what an ERP is expected to do. Businesses running finance and operations on legacy infrastructure are now asking whether their systems can support capabilities their competitors are already using. For GP, and for most on-premises ERP systems, the honest answer is that they can’t.
Key takeaways:
- AI capabilities are cloud-native by architecture. They cannot run on an on-premises ERP, regardless of how it’s extended.
- Microsoft Dynamics 365 Business Central is Microsoft’s committed path forward for businesses that need a capable ERP with AI built in. Microsoft Copilot ships with every update, automatically, to customers already in the cloud. Businesses on GP receive maintenance.
- The cost case for migration is real, but it’s no longer the primary argument. The primary argument is competitive access to AI capability that compounds over time.
- Choosing between Business Central and Microsoft Dynamics 365 Finance and Operations matters. They serve different organisational profiles, and most partners aren’t positioned to advise honestly across both.
The infrastructure gap that AI is exposing
Conversations about AI adoption tend to focus on tools, processes, and vendors. Infrastructure sits further down the agenda, if it’s there at all.
But that question is the one that determines whether the others are even answerable.
By now, there is a long list of AI capabilities that every enterprise needs: Copilot functionality, predictive analytics, automated reconciliation, natural language querying of financial data. These are all cloud-native by architecture. They’re built on platforms that update continuously and connect to other systems through APIs. To function, they need live access to clean, structured data in a cloud environment they can reach.
Local databases are a roadblock to AI-readiness
GP holds data in a local database. To make that data usable for AI, someone has to extract it, clean it, and push it somewhere accessible. At that point, it’s probably no longer live. The answer any AI tool produces from that data is already out of date before anyone reads it.
This is a real problem for almost any legacy or on-premises ERP. The architecture simply wasn’t designed for what businesses are now asking it to support. Legacy systems can be extended with middleware and custom integrations, but those workarounds add cost and latency, and they still don’t deliver what a cloud-native deployment provides as standard.
How does this change with Business Central?
Microsoft’s committed path for businesses that want both a capable ERP and AI functionality is Dynamics 365 Business Central. Copilot is built into the platform. It ships with new capabilities in every major update, automatically, to customers already in the cloud. Businesses on GP don’t receive any of those updates.
The practical difference shows up in how work gets done. A finance manager on Business Central can:
- Query accounts payable data in plain language
- Run forecasts without building them manually,
- and get anomaly alerts before they become problems.
Building equivalent functionality on GP requires Power Automate, custom connectors, and ongoing consultant involvement. Business Central makes the same outcomes achievable with no-code workflows that teams can configure themselves.
The gap between those two operating realities is going to widen. The pace of AI development inside cloud ERP platforms is accelerating, and every update increases the distance between businesses that have made the move and businesses that haven’t.
The cost argument has become secondary
There is a financial case for migration that stands on its own. Forrester’s Total Economic Impact study on Dynamics 365 Business Central modelled a composite organisation that achieved more than a 10% reduction in total cost of ownership, with over $170,000 in present-value savings from retiring legacy systems and maintenance overhead. Those figures are risk-adjusted, so they’re deliberately conservative relative to what interviewed customers actually reported.
Hardware refresh cycles, manual upgrades, patch management, backup infrastructure: the costs of staying on-premises are real, but they’re distributed across time and across teams in a way that makes them easy to underestimate until you add them up.
That said, the infrastructure savings are a supporting argument now, not the lead one. The primary question is strategic: what can your business do, decide, and build when your ERP is connected to the cloud? The Forrester numbers help justify the investment. They don’t capture the cost of falling behind on AI capability while competitors accumulate it.
Choosing the right ERP migration partner matters as much as choosing the right platform
Business Central is the right destination for most businesses currently on GP. But it’s not the only option, and the decision deserves more than a default assumption.
Microsoft’s current ERP stack includes both Business Central and Dynamics 365 Finance and Operations. They serve different organisational profiles:
- Business Central is for businesses that need a capable, manageable platform
- F&O is more suitable for larger, more complex operations with requirements around advanced manufacturing, global compliance, or multi-entity finance.
Very few partners are positioned to advise honestly across both, because most have deep expertise in one and limited exposure to the other.
Braintree works across all three: GP, Business Central, and Finance and Operations. The advice we give on which platform to move to is based on your situation, not on which migration we’re most equipped to deliver. It also means that when we assess a GP environment, we’re doing it with full knowledge of what the destination systems require.
- What the migration will involve: GP’s table structure doesn’t map cleanly to BC’s data model, and customisations that accumulated over years need to be inventoried and evaluated individually before anything moves. Some have direct BC equivalents, others need to be rebuilt. Some may turn out to be redundant against functionality BC handles natively.
- Where the complexity lives: Chart of accounts structures, multi-entity setups, and intercompany processes that worked in GP frequently surface as re-architecture decisions in BC rather than straightforward migration tasks. These are the workstreams that tend to expand scope and timeline when they’re not identified early.
- Planning without disruption: Go-live timing, parallel running duration, and end-user readiness are the three variables that most determine whether a cutover is stable. They’re often treated as late-stage logistics rather than decisions that need to be made at the start of the project.
“Later” keeps getting more expensive
Every year a business stays on GP, its competitors on cloud platforms accumulate advantages: AI capability, more efficient operations, data infrastructure that yields more advantages over time. Migrations done under pressure are more disruptive and more expensive than migrations with lead time.
Staying on-premises was a reasonable position while cloud platforms were still maturing. It bought stability. That calculation has changed, and most businesses already sense it. The harder question is whether it’s showing up in current planning, or whether it’s still in the “we’ll deal with it later” column.
The conversation to have first
Braintree’s Dynamics Discovery Assessment is designed for businesses that want to talk about Business Central or F&O with a partner fluent in both. It’s a structured conversation about your current environment and it produces a realistic picture of the path you need to take.
