Automated purchase orders means two quite different things and the difference matters commercially. In the narrow case an order is raised, approved and sent with no human involvement at all. In the wide case a person still decides what to buy and the software removes the typing, the routing and the chasing. Most spend falls in the wide case.
The narrow case: reorder against agreed terms
Stock below a level, contracted price, known supplier, predictable lead time. Here the order genuinely raises itself and the control lives in the agreement rather than in each order. It works well and applies to a smaller share of most organisations' spend than the phrase implies, so measure yours before buying on the promise.
The wide case: automatic paperwork
Numbering, routing for approval, issuing, chasing acknowledgement, tracking receipts and flagging overdue deliveries, with a person choosing what to buy. This is the majority case and it is where most of the elapsed time and most of the errors are, so it is worth having even though it sounds less impressive.
What must be true before either
Current prices, clean supplier records and somebody reviewing what the automation produced. Automation multiplies existing discipline in both directions: an automatic order against a stale price list commits you at the wrong price faster than a person would have, and nobody notices until the invoice arrives.
Questions people ask about automated purchase orders
What share of our orders could be automatic?
Look at repeat purchasing against agreed prices with predictable demand. For most organisations it is a minority.
Is automatic ordering risky?
Only in proportion to how stale your price agreements are. Review dates on agreements matter more once orders raise themselves.
Does it help payables?
Yes, because it produces orders and receipts for the match. Most of payables' benefit from ordering automation is upstream discipline.