Legacy ERP infrastructure is becoming an active constraint on the ability of companies to invest in and benefit from AI, according to Courtney Hounsell, Client Experience Manager at Braintree. Migration is not modernisation, Hounsell argues, and many businesses have confused the two — leaving them without the foundations needed for AI to work effectively.
Legacy ERP hampers AI progress.
Cloud migration has often been sold as digital modernisation, but as Hounsell notes, it “only changes the address of the system.” Real modernisation alters architecture, data models, integration logic, and the capacity to generate value. When businesses conflate the two, they end up paying cloud prices for on‑premises problems and arrive in the AI era without the necessary groundwork.
Related: How to Choose the Best Cash Flow Management Software for Growing SMBs
The gap between technology investment and business outcomes creates its own set of problems. Invisible data, constrained AI capacity, and system incoherence tend to fall through the cracks, leaving firms behind on the adoption curve. Capabilities that define competitive operations in 2026 — such as predictive analytics, natural language querying, automated reconciliation, and intelligent workflow routing — are cloud‑native by design and cannot run on a legacy on‑premises ERP.
AI adoption is rising, but measurable impact remains rare
A recent report on the State of AI in 2025 surveyed nearly 2,000 executives across 105 countries. It found that 88% of companies have started using AI in at least one area of the business, up from 78% in 2024. Yet only 39% could attribute any measurable EBIT impact to AI. As the analysis notes: “Meaningful enterprise‑wide bottom‑line impact from the use of AI continues to be rare.”
Firms that are extracting value share one structural characteristic — the AI can see their core systems. That’s where on‑premises ERP platforms struggle. They hold data in local databases, requiring extraction, cleaning, and pushing to an accessible location before AI can use it. This process slows the entire AI insights value chain because the data is no longer live, and the insights become out of date.
Related: Boosting Your Bottom Line: The Benefits of Credit Card Payment Processing Services for Your Business
Middleware and custom integrations don’t fully solve the problem. These workarounds often cost more and introduce latencies that still fall short of what a cloud‑native deployment provides as standard. Without real‑time data pipelines feeding into core systems, AI deployments remain siloed and limited.
There is a clear financial case beyond AI
According to the Forrester TEI study on Trends 365 Business Central, companies moving into ERP modernisation can see a potential return on investment of 265%. Measurable productivity gains appear 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.
When ERP systems remain on‑prem, the real question isn’t what gaps remain or what siloes limit collaboration. It is what cannot be done. That question becomes more urgent with each update cycle. ERP platforms benefit from automatic updates and refreshes, now bringing new AI capabilities with each cycle, while on‑premises platforms receive only maintenance.
