Days payable outstanding (DPO) is the average number of days a business takes to pay its suppliers after receiving their invoices. It is calculated as average trade payables divided by cost of sales, multiplied by the days in the period. A higher figure means cash stays in the business longer; too high, and supplier relationships suffer.
Why does DPO matter once agents act?
DPO is a policy outcome. A treasurer decides whether to pay on the due date, pay early for a discount, or hold to the last day of terms. When an AI agent runs invoice processing and payment scheduling, that policy has to be written into the agent's bounds, or the agent sets it by accident.
The common accident is speed. An agent that clears invoices the day they arrive and schedules them for the next payment run pays suppliers early. Processing faster is good; paying faster is a decision about working capital that nobody took. The reverse also happens: an agent holding invoices with unresolved queries can stretch payments past terms and into late payment. Both show up in DPO before anyone notices them elsewhere, which is one reason it belongs on the dashboard for AI agents for finance.
How is DPO calculated?
The standard formula:
DPO = (average trade payables ÷ cost of sales) × number of days in the period
- Average trade payables: opening plus closing supplier balances, divided by two.
- Cost of sales: the cost base the payables relate to. Some teams use total purchases instead, which is more accurate when large costs are not bought on credit.
- Days: 365 for a year, 90 for a quarter.
Use the same definitions every period. A change of formula moves DPO more than any change in behaviour.
What does that look like in an example?
A manufacturer has annual cost of sales of 3,650,000, or 10,000 a day, and average trade payables of 450,000. Its DPO is 45 days, in line with its standard terms.
It introduces an agent that reads, codes and matches supplier invoices within hours, and payment runs pick up everything approved. Within a quarter the payables balance falls to 300,000 and DPO to 30 days. Nobody changed payment policy. The fix is a scheduling rule in the agent's bounds: approved invoices are scheduled for the payment run before their due date, not the next one. The agent keeps its processing speed and DPO returns to terms. The connection between agent and payment tool is set out in how to connect an AI agent to your accounts payable automation tool.
How is DPO different from days sales outstanding?
They are mirror images. DPO measures how long you take to pay suppliers; days sales outstanding measures how long customers take to pay you. Together with inventory days they make up the cash conversion cycle: inventory days plus DSO minus DPO. A business can improve that cycle by collecting faster or paying slower, and the two levers carry different risks.
DPO is also not the same as a measure of late payment. A DPO of 60 is fine if terms are 60 days. Track payments made after due date separately, because an agent that improves DPO by paying late has not improved anything.
Which related terms should you know?
- Accounts payable: the amounts owed to suppliers that DPO is built from.
- Payment terms: the agreed days to pay, which set the benchmark for DPO.
- Early payment discount: a reduction for paying before the due date, which lowers DPO by design.
Questions people ask
Should an AI agent decide when to pay suppliers to manage DPO?
The agent should apply a payment policy, not set one. Treasury decides the rule, such as pay on the due date unless a discount is worth taking. The agent schedules within it, and anything outside the rule, such as paying a strategic supplier early, is held for a person.
Is a higher DPO always better?
No. Beyond agreed terms it means paying late, which damages supplier trust and can breach contracts or local late-payment rules. The useful target is paying on terms, consistently, with early payment as a deliberate choice.