A rule engine is a brilliant idea that has aged poorly. It was supposed to give business teams control over policy. Instead, it produced a generation of systems that only IT can edit.
Why this happened
Rule engines turn policy into code. Even when the syntax looks declarative, it ends up requiring engineering judgement — versions, dependencies, regression testing, deployments. Business owners stopped touching them; IT became the bottleneck.
Plain language is the new contract
Large language models can read and apply policy expressed the way a human would write it. The business owner describes what should happen. The system executes it. There is no translation step.
Why this is structurally different
- A policy change takes hours, not weeks.
- The owner of the rule is the owner of the process.
- Audit trails are richer — you have the rule, the data, and the reasoning.
The shortest distance between a policy and its execution is a sentence.