AI Prompt for a Grant Budget Justification
This AI prompt for a grant budget justification drafts the budget narrative from your finished, finance-approved budget. For each line you supply the amount and how it was calculated. The AI states the calculation, ties the cost to a named project activity and follows the funder's categories. It never changes a figure. It rechecks the arithmetic and reports any line that does not add up, has no stated basis or has no linked activity.
The prompt
You are helping me write a grant budget justification (budget narrative). Use only the information I provide. Project: [PROJECT NAME AND ONE-LINE DESCRIPTION] Funding period: [PERIOD] Funder's budget categories and instructions: [PASTE BUDGET INSTRUCTIONS] Project activities: [LIST NAMED ACTIVITIES] Approved total request: [TOTAL] Approved budget lines. For each line: category, item, exact amount, how it was calculated (quantity, rate, period), the reason for the cost, and the activity it supports: [PASTE BUDGET LINES] Step 1. Check before you write. - Recalculate every line from the basis I gave and compare it with the stated amount. - Add the lines and compare the sum with the approved total. - List every line that (a) does not match its own calculation, (b) has no stated basis, or (c) has no linked activity. Report these. Do not correct them. Step 2. Write a short justification for each line, organized under the funder's categories in the funder's order. For each: - State the calculation exactly as I supplied it. - Say what the cost is for and which activity it supports, using my stated reason. - Keep to two to four sentences. Rules: - Do not change, round, estimate, combine or split any figure. - You may calculate only to check my arithmetic, never to create or replace a number. - If something is missing or inconsistent, write "Needs confirmation" on that line. Do not invent a reason, an activity, an assumption or a corrected amount. - Do not state that a cost is allowable or that the budget complies with any rule. Return, in this order: 1. Issues needing confirmation 2. The budget justification 3. A table listing each line, the amount in my budget and the amount in your narrative, so I can confirm they match
Written for ChatGPT, Claude, Gemini and similar AI chat tools. Replace each bracketed placeholder with your own details before you send it. Results vary by tool and version, so always check the output.
What to fill in
- [PASTE BUDGET INSTRUCTIONS]
- The funder's budget categories, required format, any length limit and the terms they use, copied from the announcement or budget form.
- [LIST NAMED ACTIVITIES]
- The project's activities with the same names used in your narrative, so each cost can be tied to one of them.
- [TOTAL]
- The total request from your approved budget. The AI checks that the lines add up to it.
- [PASTE BUDGET LINES]
- Every line from the finance-approved budget, with the amount and the basis for it: quantity, unit cost or rate, and period. Use role titles for personnel, not names. "Materials for 12 participant workshops" ties a cost to the work better than "project supplies."
Why this prompt works
A budget justification explains numbers that already exist. The risk in using AI for it is that the tool alters a number, or invents a reason for one. Every part of this prompt is aimed at those two risks.
- Checking comes before writing. Step 1 has the AI recalculate each line and the total and report mismatches. Errors surface before they are wrapped in fluent prose.
- Report, do not correct. If a line does not add up, the AI tells you. It does not decide which number is right. That decision belongs to finance.
- Figures are locked. No rounding, combining or estimating. The only arithmetic allowed is the check.
- Structured inputs. Each line carries its amount, basis, reason and activity. The justification is assembled from those four parts, so a missing part is obvious.
- "Needs confirmation." Where the AI would otherwise invent a rationale, it marks the line instead.
- A match table. The final table puts your amount beside the AI's for every line, which is the quickest way to confirm nothing drifted.
- No allowability claims. The AI is barred from saying a cost is allowable, because it has no basis for that judgment.
AI tools do make arithmetic mistakes, so treat its check as a second look. It does not replace your own.
Why it is appropriate to use
Use this once the budget itself is final. Building the budget is finance and program work: real quotes, real salaries, real rates. Explaining it clearly, line by line, in the funder's format is the writing task that follows, and it is repetitive enough that AI is a real help.
That order is what makes this an appropriate use. Finance staff own the numbers. The program lead confirms that each cost supports the activity named. The AI drafts the wording between them.
Do not use the prompt to set figures, choose salary levels, estimate prices or decide what a funder will pay for. Do not rely on it to interpret a funder's budget rules, indirect cost terms or equipment definitions. Those answers come from the funder's instructions and from qualified staff.
Keep payroll records, salaries tied to named people, bank details and internal financial documents out of AI tools. Use role titles and approved line amounts, and check your organization's data rules and the funder's AI policy first. For how to build the budget itself, see grant budget fundamentals.
How to check and revise the output
- Start with the issues list. Have finance resolve every arithmetic mismatch and missing basis in the source budget.
- Check the match table yourself, line by line, against the approved budget. Then recalculate the total independently.
- Have the program lead confirm each activity link and each stated reason.
- Remove anything invented: a detail, a vendor, a justification you did not supply.
- Confirm the categories, order and labels match the funder's budget form exactly.
- If the budget changes, change the source first, then redo the affected lines. A wording edit must never become a budget edit.
Frequently asked questions
- Is a budget narrative different from a budget justification?
- The terms are generally used for the same thing: the written explanation that accompanies a grant budget and shows how each cost was calculated and why it is needed. Use the term and format the funder uses.
- What if the AI finds a calculation error?
- Leave the figure as it is until your finance staff have looked at it. If they confirm an error, correct the approved budget first, then update the narrative to match.
- Can AI create my grant budget for me?
- It should not. An AI tool does not know your salaries, your vendors' prices or your funder's rules, so any figure it produces is a guess. Build the budget from real costs with your finance staff, then use AI to help explain it.
- Can I use the output without editing?
- No. Treat it as a first draft. Confirm every figure, activity link and funder requirement, and resolve every line marked as needing confirmation before you submit.