AI has one major superpower to bring to the world of ERP — and it's not the obvious one about agents. It's something much more profound. First, a little history lesson.
All enterprise SaaS applications are underpinned by a simple concept: a single code line. This provided the compelling reason for customers to migrate from on-premise, client-server applications to the cloud. The cost to maintain your own private code line — with all its incipient customisation was nose-bleedingly expensive for most customers!
The single code line that was constantly updated and maintained by the SaaS vendor was the siren call that couldn't be resisted.
However, a single code line has a dark underbelly: customer configuration!
In order to implement a SaaS solution for a specific customer, a whole ton of configuration has to be created and managed: workflow definitions, custom fields, security configuration, report definitions and so on. In the case of very large ERP suites these customisation projects run into the $Ms! However, that is not the worst part. The real sting is when you want to modify the configuration. One recent example highlights the issue. A large customer of a HCM SaaS suite wanted to turn on time & attendance for the dozen or so people in Spain. The price ran to $Ms again! The issue is that the configurations are so complex and opaque that they are hard to reason about.
With AI now showing the way about what can be achieved with code generation the world is being inverted. Future ERPs will be fully customised and generated applications — but this time the power of AI makes the management and update of these generated applications a much more tractable problem. Furthermore it offers enterprise customers something really new: a fully customised application that meets their specific needs — at a fraction of the cost.
