The example is made up. Hillside Literacy Project is a made-up nonprofit that runs after-school reading programs, and its donors, campaigns, foundations and every number are made up. The board will use the report to decide whether to hire a second reading coordinator, so a total that is off by a pledge or two matters. The practice copy of ChatGPT needs no account, and no file leaves your computer. This is our own lesson, not an official OpenAI course.
How does the analysis work, step by step?
The lesson has 26 steps in five short chapters:
- Upload the file. Click the + on the left of the message box, choose Add photos & files, pick hillside-gifts-jan-jun-2026.csv and click Open. An Excel file works the same way.
- Check the data first. Click Paste request and Send. The request asks what each column means, the row count, the date range and any duplicates, blanks or rows that may not be money received, with no insights yet. Then open Problems found.
- Clean it your way. Click Paste fix: remove the second copy of each synced gift and mark the pledges Pledge - not received. Open the cleaned file, then close the panel with the X.
- Ask your question. Click Paste question for gifts received by month, split by source, and January to March against April to June.
- Chart and takeaways. Ask for a line chart, open it, click download, then ask for three takeaways and click Copy.
The export is a CSV file: one row per gift or pledge.
What did the check find?
The file has 1,212 rows and 7 columns: gift_date, donor_id, donor_type, campaign, gift_type, transaction_id and amount. Names and contact details were removed; each donor has an ID. The dates run January 2 to June 30, 2026, with no blank dates or amounts.
Two problems hide in it. Five online gifts appear twice with the same donor, date, transaction ID and amount, because the giving platform synced them twice. Left in, $1,240 would count twice. And 14 rows are pledges, $18,750 in total: promises to give that may not be money received yet. Counted as gifts, the board would see money the nonprofit doesn't have.
Who decides what counts as a gift?
You do, not ChatGPT. The fix tells it the duplicates are sync errors and the pledges are promises. It removes the second copy of each duplicate and changes the pledges' gift_type to "Pledge - not received", so they stay in the file but out of every total. The cleaned data has 1,207 rows, and 1,193 of them are gifts received. ChatGPT gives you the cleaned file to keep for the auditor. Removing duplicates is the easy part; deciding how pledges are treated is a policy call.
Try it free, step by step
Analyze a Donor Gift Export with ChatGPT: Board-Ready Giving Numbers for a Nonprofit. Upload your donor database’s gift export to ChatGPT, catch gifts that synced twice and pledges that were never received before anything is added up, then get giving by source, a chart and three… You do every step yourself in a practice copy of ChatGPT, it checks your work as you go, and “Do it for me” finishes any step you get stuck on. Free, and nothing touches your real accounts.
Start the free lessonWhat did the numbers show?
$157,765 received from January to June. Every source grew in the second quarter. The three takeaways for the board:
- The Spring Appeal worked. One-time gifts jumped from $8,400 in March to $21,600 in April, the month the appeal went out.
- Monthly donors are the steadiest money. Monthly giving grew every month, from $3,150 to $3,725, 13.0% of everything received.
- A third of giving came in two checks. Foundations gave $55,000, 34.9% of the total, in two grant payments, so a late or smaller grant would leave a large gap.
Key takeaways
- Remove names and contact details before you upload. A donor ID is enough.
- Ask for a check of the file before any totals or insights.
- Decide yourself how to handle duplicates and pledges.
- Split giving by source: one big total hides what drives it.
- Tie the totals to recorded deposits before the board sees them.
Frequently asked questions
- Can ChatGPT read an Excel file instead of a CSV?
Yes. The lesson says an Excel file works the same way.
- How does ChatGPT add up the rows?
For questions like this, ChatGPT usually writes and runs a little Python code, so the totals are worked out from the file itself.
- Why mark pledges instead of deleting them?
Marking keeps them in the file, so the report can show the board what is pledged and not received yet, separately from money in the bank.
- Why ask for a quarter-to-quarter comparison?
Because March and June look big only because two foundation checks came in. Splitting by source and comparing January to March with April to June shows what one big total hides.
- What kind of chart does the lesson make?
A line chart by month, one line each for one-time gifts, monthly donors, foundations and businesses, with a clear title. You open it full size and click the download icon to save it as a picture for the report.
- What did ChatGPT say to double-check?
List the 14 pledges separately as "pledged, not received", ask your finance team or bookkeeper whether the June foundation payment is for this year's grant or an early payment of next year's, and check that the totals match the deposits they recorded.
- Is it safe to upload donor data?
Remove names, emails, addresses and card details first, use a business ChatGPT plan that doesn't train on your data, and follow your privacy policy and any promises you made to donors. The PII entry explains what counts as personal data.
- Where can I read about the common slip-ups?
See ChatGPT donor gift analysis mistakes and is it safe to paste client data into ChatGPT. The base ChatGPT spreadsheet lesson uses the same steps for other data, and the nonprofit board meeting recap lesson helps after the meeting.
