Legacy ERP infrastructure is becoming an active constraint on the company’s ability to invest in, and benefit from, AI’s capabilities, writes Courtney Hounsell, client experience manager at Braintree.
Migration is not modernisation. Unfortunately, many companies have merged the two, and the result is a gap that’s creating incoherence and a lack of visibility, constraining their ability to benefit from AI and its capabilities.
Cloud migration has been sold as digital modernisation, but this only changes the address of the system. Real modernisation changes architecture, data models, integration logic, and the capacity to generate value. Companies want their modernisation projects to actively reduce other costs, show long-term ROI, and empower their staff with systems that work faster and smarter. So, when migration and modernisation are confused, companies end up paying cloud prices for on-prem problems and arrive in the AI era without the foundations required to ensure AI can function effectively.
This widening gap between technology investment and business outcomes creates its own set of problems. Invisible data and constrained AI capacity and system incoherence tend to fall in the cracks left by this gap and leave companies behind on the curve of adoption. The capabilities that define how competitive companies are operating in 2026, such as predictive analytics, natural language querying, automated reconciliation, and intelligent workflow routing, are all cloud-native by architecture and can’t run on a legacy on-premises ERP.
In a recent McKinsey report on the State of AI in 2025, the survey across nearly 2,000 executives in 105 countries found that 88% of companies have started to use AI in at least one area of the business, an increase from 78% in 2024. However, only 39% are able to attribute any measurable EBIT impact to AI.
As McKinsey says:
‘Meaningful enterprise-wide bottom-line impact from the use of AI continues to be rare.’ Companies that are extracting value share one structural characteristic, the AI can see their core systems.
On-premises ERP systems are struggling uphill within the AI economy. They hold their data in local databases, and to make this usable for AI, it has to be extracted, cleaned and pushed somewhere accessible. And this entire process slows down the entire
AI insights value chain because the data is no longer live, and the insights it produces are out of date. It’s a challenge felt by many companies sitting on ERP platforms that aren’t quite agile and accessible enough for the onslaught that is AI, and it also can’t be completely resolved with middleware or custom integrations.
These workarounds can often cost the company more money and introduce latencies that still don’t deliver what a cloud-native deployment can provide as standard. Without real-time data pipelines feeding into core systems, AI deployments remain siloed and limited.
There is a financial case for moving from legacy ERP to a more agile and modernised cloud deployed system that sits outside the AI discussion as well. According to the Forrester TEI study on Dynamics 365 Business Central, companies moving into ERP modernisation can also see a potential return on investment of 265% with measurable productivity gains across operations (12.5%), sales (15%) and finance (15.6%).
Hardware refresh cycles, manual upgrades, patch management and backup infrastructure carry costs that are easy to underestimate because they’re distributed across time and teams.
And as companies keep their ERP systems on-prem, the real question isn’t what gaps are left or the siloes that limit collaboration, but rather what you can’t do. And that question is becoming more urgent with each update cycle.
ERP systems benefit from automatic updates and refreshes, now bringing new AI capabilities with each update cycle, where on-premises platforms are only receiving maintenance. It’s a sticky space to be in, especially for companies that are trying to minimise expenditure and tighten budgets.
Fortunately, ERP platforms and integrations have modernised alongside the architecture. ERP migration partners work across the full Microsoft stack, from legacy on-premises environments through to Dynamics 365 Business Central and Finance and Operations.
Business Central has become a solid fit for companies wanting a capable and manageable platform built for growth, while Finance and Operations slots into larger and more complex companies.
Getting the tech right before the migration determines whether or not the project delivers on its promise or simply moves the problem to a new postcode. It also addresses the cost problem, optimising migration to budget expectations and a clearly defined ROI.
The same Forrester study found that companies that have made this move correctly have avoided more than $30,000 in third-party consulting fees annually. It’s a saving that comes from reduced complexity and streamlined optimisation through the cloud, and is why the quality of the migration partner is as important as choosing the right ERP.
Modernisation is not migration, but the right migration is definitely a step towards modernisation and removing the barriers inhibiting access to next-generation AI capabilities.
