Attempts to streamline accounts payable usually start with a solution and work backwards, which is why so many of them improve the part that was already cheap. A week of measurement, costing an hour a day, reorders most improvement plans and frequently identifies a free fix that outperforms the purchase somebody was about to make.
What to measure in that week
For each invoice: how it arrived, whether it had a purchase order and a receipt behind it, how many minutes of human attention it took, and how many days elapsed from arrival to approval. Four fields, recorded as you go. The gap between minutes and days is usually the finding that changes the plan.
What the measurement usually shows
That most invoices take almost no time and a minority take most of it. That the minority are dominated by missing orders and missing receipts. And that elapsed time is dominated by waiting rather than work. Those three facts point at upstream discipline and approval routing rather than at capture, which is where most projects start.
The order to act in
Consolidate arrival channels, which is free. Fix receipt recording at the delivery point, which is free. Route approvals to named people with cover and escalation, which is configuration. Then, and only then, consider capture, matching and payment tooling, each sized by what the measurement showed.
Questions people ask about streamline accounts payable
Is a week enough?
Usually, unless volumes are very low or highly seasonal. You are looking for the shape of the distribution rather than precise averages.
What if we cannot spare the effort?
Sample fifty invoices instead of all of them. A sample answers the same questions well enough to direct the work.
What is the most common surprise?
How much of the elapsed time is waiting, and how little of the total effort the majority of invoices consume.