Microsoft Copilot · Microsoft
Fix invoice capture errors before they reach finance
Asurion's IT team put a Copilot Studio agent inside its invoice capture flow to repair common data errors, such as messy purchase order numbers, before invoices post to Dynamics 365 Finance.
What they did
When invoice capture flags a record in an error state, a Power Automate flow sends the details to the agent. The agent follows set instructions, for example trimming item numbers to match the right purchase order or stripping extra text from PO numbers. A follow-up flow reprocesses the corrected invoice into Dynamics 365 Finance.
How it went
The company says touchless invoice processing in the Americas rose from under 20% to over 70%, and the first proof points were built in about a week. It also kept its existing capture setup instead of switching OCR vendors.
Worth knowing
This is a vendor-published story with no baseline method shown. Large PDFs, exception-heavy vendors and rollout to Asia Pacific and Europe were still in progress, and human review stays at key points.
Try it with Microsoft Copilot
You receive invoice records that failed capture in <your AP system>. For each one, check the PO number and item numbers against <your open purchase order list>, clean up formatting problems such as extra text or wrong lengths, and return the corrected fields with a short note on what you changed. If you cannot find a confident match, mark it for human review instead of guessing.
We wrote this prompt from the case. Replace the parts in angle brackets with your own details.
What this label means: Vendor says
The maker describes it on its own site or launch post. It shows what the product is meant to do, with no independent result.
Real cases of AI agents at work, from every major agent, each linked to where it was first reported. Rewritten in our own words, most with a prompt you can try. These are not Wize customers.