Daily Coupa invoice exception triage for your AP team

Every weekday at 7am, triage my Coupa invoice exception queue so my accounts payable team walks in knowing exactly what to fix first, instead of clicking through held invoices one at a time. Use a cron trigger set to weekdays at 7am. Coupa does not send outgoing webhooks through this integration and is not available as a polling source, so cron is the correct and only valid trigger here.

Start by pulling the exception queue from Coupa with the List Invoices operation. I only want invoices sitting in exception states rather than the whole ledger, so filter for statuses such as on hold, disputed, and pending review. Note that Coupa list endpoints return at most 50 records per call and support offset based pagination only, with no cursor and no way to raise the page size. Loop by increasing the offset in increments of 50 and stop when a call returns fewer than 50 records. Apply the same pagination rule to every Coupa list call in this workflow.

For each exception invoice, gather the context needed to judge it. Fetch the linked purchase order with the List Purchase Orders operation, and fetch the vendor record with the List Suppliers operation. Then compare the invoiced unit price, quantity, tax, and currency against what was actually ordered on the purchase order.

Important scope limit: this is a two way match between the invoice and the purchase order only. There is no receiving or inventory receipt list operation available, so do not attempt to verify goods receipts and do not describe the result as a three way match anywhere in the output.

Diagnose the likely cause of each exception in plain language that an AP clerk can act on. The common cases are: unit price invoiced above the price on the order, quantity billed above quantity ordered, no matching purchase order at all, a tax discrepancy, or a currency mismatch. Be specific and quantitative rather than vague. For example, say "Billed at $62.00 per unit against $48.00 on PO 4471, across 200 units" rather than "price mismatch". Where the supplier record explains or contradicts the variance, such as a tax rate or default currency on file, mention it.

Apply a tolerance rule so small variances do not drown out the real problems. Treat a variance as low risk when it falls under the tolerance threshold, defaulting to 50 dollars or 2 percent of the invoice line value, whichever comes out smaller in dollar terms. Group every low risk item into a single summary line giving the count and the combined value, rather than listing them individually. Expose both the dollar and the percentage threshold as settings at the top of the workflow so I can tune them without editing the logic.

Assign each remaining exception a priority based on two factors: the value at risk, meaning the absolute dollar variance between what was invoiced and what was ordered, and how long the invoice has been held, meaning the number of days it has been sitting in an exception status. Items that are both high value and long held should rank at the top. Rank the full list so the team can work straight down it.

Post the ranked triage brief to my accounts payable Slack channel using Slack's Send a Message operation. Lead with the headline numbers, the count of held invoices and the total value at risk. Then give the ranked list, and for each exception include the invoice number, supplier name, value at risk, days held, the diagnosed cause in plain language, and a suggested next step. Close with the single grouped low risk line. If there are no exceptions at all that morning, post a short all clear message instead of an empty brief, so the team knows the check ran.

Append one row per exception to my Google Sheets tracker using the Append Values operation, so the team can watch exception aging and spot repeat offender suppliers over time. Include columns for date triaged, invoice number, supplier name, invoice amount, purchase order number, variance amount, variance percentage, diagnosed cause, priority, and days held. Append rather than overwrite so history accumulates. Log every exception individually, including the ones collapsed into the low risk line in Slack, because the grouping is a presentation choice for the brief and the tracker should keep full detail for trend analysis.