Just a year ago, many companies were wondering how to implement generative AI. Today, it is increasingly becoming clear that the question should be phrased differently: are our ERP and CRM systems even ready for AI to deliver real value? It is the quality of the data, the architecture of the systems and the degree to which they are integrated that are increasingly determining the success of AI projects, rather than the capabilities of the models themselves.
This is changing the way we think about modernising business systems. ERP and CRM are no longer seen as costly infrastructure projects. They are becoming the foundation for process automation and the use of AI on a larger scale.
AI does not operate in a vacuum. To prepare a quote for a customer, forecast demand, plan production or automate order processing, it needs access to data stored precisely within ERP and CRM systems. What is more, it is within these systems that AI subsequently carries out its tasks – updating information, triggering processes or assigning tasks for execution. This means that AI does not replace transactional systems, but rather enhances their importance. The better organised these systems are, the greater the value AI can deliver.
The biggest problem, however, turns out to be not the age of the systems, but the quality of the data. Over the years, many organisations have developed their ERP and CRM systems independently of one another. This has led to a succession of integrations, local modifications, duplicate customer records and inconsistent product catalogues. Humans can often cope with such chaos because they understand the context and can spot errors. AI lacks this ability. If it receives conflicting information, it generates conflicting responses or takes the wrong course of action. From a business perspective, this means that artificial intelligence does not fix data problems – it merely exposes them more quickly and on a larger scale.
This is precisely why more and more organisations are reordering their investment priorities. Instead of starting with yet another AI pilot, they first organise their data, simplify their architecture and integrate their systems. Only on this foundation do they build solutions utilising generative AI or AI agents. In practice, this means a much greater chance of moving from isolated experiments to process automation that delivers tangible business benefits.
However, this does not mean that the entire ERP or CRM system needs to be replaced. Fewer and fewer organisations are opting for long-term ‘rip and replace’ projects, which involve high risk and often paralyse operational activities. Instead, modernisation takes place in stages. Companies are organising their reference data, making it available via APIs, simplifying integrations and gradually migrating selected areas to a more modern architecture. This enables them to deploy AI where it delivers the greatest value, without having to overhaul the entire IT environment.
This approach is also changing the role of the CIO. Until recently, the success of modernisation was measured by the timely completion of a migration or by reducing system maintenance costs. Today, preparing the organisation to utilise AI is becoming more important. This encompasses not only technology, but also consistent data, standardised processes and an architecture that allows information to be securely shared with other services and AI agents.
As a result, ERP and CRM modernisation ceases to be an end in itself. It becomes an investment that determines whether AI will be yet another costly experiment or a tool that genuinely boosts the company’s productivity. Organisations that treat the streamlining of their systems as the first stage of their AI strategy will have a much better chance of successfully automating processes than those that attempt to build intelligent solutions on an inconsistent foundation. This is precisely why the greatest investment in AI today is a modern, integrated and well-managed ERP and CRM system.
