What is the difference between automation and AI?

It matters more than it sounds, because one of these is significantly cheaper and more predictable than the other.

Automation

Rules. When this happens, do that. It runs identically every time, it is cheap, and when it breaks it breaks loudly and obviously.

If a task can be written as a flowchart with no "it depends" boxes, this is what you want. Most of the value in most businesses is here, and it is deeply unfashionable.

AI

Judgement over messy input. Reading an email and working out what it is about. Summarising a document. Extracting data from a supplier invoice that has a different layout every time.

It handles variation that would need a thousand rules. It costs more per operation, and it is occasionally wrong in ways that look right.

Why the distinction earns you money

A lot of what gets sold as AI is a rule with a language model bolted on for glamour. You pay more, you get less reliability, and the flowchart would have been fine.

The reverse also happens — a business tries to write rules for something genuinely variable, ends up with four hundred exceptions, and abandons it.

The practical test

Could a new starter do this correctly from a written checklist? Then automate it with rules. Does it need someone to read something and decide? That is where AI earns its cost. Most real systems are mostly rules with AI at the two or three points where judgement is needed.