Use case #27 · CPA
Nonprofit controller routes scanned mail through an LLM summary into Jira tickets
- Close/bookkeeping
- Intermediate
- Microsoft Power Automate, A secure large language model (not named), Jira
- Source checked · Journal of Accountancy article
The short version
What they did
Spanbauer’s accounting team was buried in physical mail: vendor invoices, tax notices, statements. Mail is now scanned to a shared drive, its text extracted, and an LLM writes a summary; a Jira ticket is opened with the summary and the original for the team to work. He later added automations to look up the status of scanned invoices.
Result (as claimed)
“The team now tackles the mail in 20-minute sprints for ‘Mail Mondays,’ saving about four hours of total staff time per week.”
Prompt
I work on a nonprofit’s accounting team. I’ll attach [one scanned letter or invoice], or paste [its extracted text]. Help me prepare a summary for a Jira ticket so a person can handle the mail.
1. Summarize it in three short lines: who sent it, what it is, and what action it asks for by when.
2. Give separate labeled fields for vendor or agency name, document type, amount, account or notice number, and due date.
3. Add a brief list of anything a team member needs to check against the original scan.Use plain language and only what the document says. Write “unclear” for unreadable text and “not stated” for missing fields. Don’t guess a deadline or infer that an invoice is unpaid. Return text I can review before putting it in the ticket.
Do it yourself
These steps are my suggested way in, not Karl Spanbauer’s setup.
Level: Intermediate for the automated flow into a ticket system. Summarizing scanned mail in a chat is Beginner.
What you need
- A scanner or scanning app that saves PDFs to a shared folder.
- A chat assistant that reads PDFs.
- For the full version: an automation tool such as Power Automate, and a task or ticket system.
- Use a business plan of the AI tool, or sample data, for anything with client details in it.
Start simple
- Scan a week of mail into one folder.
- Paste the Prompt and attach a few scans, one at a time.
- Put the fields it pulls out into a simple tracking sheet.
- Work the sheet in one short session, as Spanbauer’s team does on Mondays.
Take it further
- Have an automation pick up each new scan, get the summary and open a task with the original attached.
- Route by type: invoices to payables, tax notices to the tax lead.
- Add a lookup that shows whether a scanned invoice is already in the books.
Check your work
- Check amounts and due dates against the scan before you pay or respond.
- Look for “unclear” fields. They should go to a person, not be filled in.
- Spot-check that every scanned item became a task, so nothing gets lost.
Before you rely on it
- This is general information about how accountants use AI, not professional guidance for your engagements.
- You stay responsible for the work under your professional standards, including review and sign-off. An AI tool’s output is a draft, not evidence.
- Before you send client data to an AI tool, read the tool’s data and training terms, and get any client consent you need. For tax return information, that includes IRC section 7216.
Source checked. The link, author, date and quote were confirmed against the source. The result itself has not been reproduced. Prompt adapted for this library, inspired by the case. Business size: Not stated (nonprofit food bank). Results are quoted as the source claimed them and never improved. Spot an error or a broken link? Email charlie@cbarmorecpa.com.