Two weeks before a funder report is due, attendance, survey scores and spending are pulled into the funder's template: calculated figures and charts, an AI-drafted narrative and a source tag on every number. The director edits and approves the draft in Google Docs.
Typical results
≈200 hrs/yr
Saved across about 10 funder reports a year
100%
Figures in each draft tagged with their source
14 days
Before the due date, when data gathering starts
Figures reflect typical results for this kind of automation.
The challenge
The program submits about 10 funder reports a year, each with its own template, period and outcome questions. The director spends around three days per report copying attendance from sign-in sheets, matching survey scores by participant and asking finance for spending by budget line. Numbers sometimes disagree between reports, and a few reports go in late.
Before and after
Hours to prepare one funder report
−83%Reports submitted after the due date per year
−100%Figure corrections requested by funders per year
−80%The automation
Two weeks before a report is due, the automation gathers the period's attendance from Google Sheets sign-ins, pre and post survey scores from the participant forms, and spending by budget line from the accounting export. Counts, outcome changes and charts are calculated from that data and placed in the funder's template, AI drafts the narrative, and every figure is tagged with its source. The director edits and approves in Google Docs, and the PDF is filed in the grant folder.
How it works
- 01
Starts 14 days before the due date
Two weeks before a funder report is due, the automation opens the report job and pulls the reporting period's data.
- 02
Data from three sources
Attendance comes from the Google Sheets sign-in records, pre and post survey scores from the participant forms, and spending by budget line from the accounting export.
- 03
Template filled, narrative drafted
Counts, outcome changes and charts are calculated for each section of the funder's template, AI drafts the narrative around them, and every figure is tagged with its source.
- 04
Director edits and approves
The program director reviews the draft in Google Docs, edits the narrative and approves it. A question about any number leads straight to its source row.
- 05
Filed and shared
The approved report is saved as a PDF in the grant folder, and the finance manager gets a copy along with the spending figures it used.
Starts from
- Attendance sign-in sheets
- Pre and post surveys
- Accounting export
Delivers
- Draft in funder template
- Approved report PDF
- Copy to finance manager
The impact
The director spends about four hours reviewing and editing each funder report instead of three days assembling it, roughly 200 hours saved across the year. Every number traces back to a sign-in row, survey or ledger line, figures match across funders, and reports go in on time.
Questions about this automation
Can it follow each funder's own report template?
Yes. Each funder's template is set up once with its sections, required figures and word limits, and the draft is built in that structure. When a funder changes its form, the template is updated and the next report follows it. Sections the data can't support are left marked for the director to write.
How do we know the AI didn't make up a number?
The AI writes the narrative, while the figures come from calculations on the source data, and each one carries a note pointing to the sheet, survey or budget line behind it. The director can click through to check, and nothing is sent to a funder until the director approves the draft.
What data does the program need to have in place?
Attendance recorded per session in Google Sheets or an attendance app, pre and post surveys collected through a form tool, and an accounting export with spending by budget line. The grant calendar with due dates and templates is added at setup. This is a multi-step build, usually rolled out in stages.