Grant writing is time-intensive and detail-heavy — perfect for AI assistance. AI won’t replace the human expertise and relationships that win grants, but it dramatically accelerates drafting and research.
What AI Can (and Can’t) Do for Grant Writing
AI excels at:
- Structuring proposals to funder specifications
- Drafting narrative sections from bullet points
- Research on funder priorities and past grants
- Editing for clarity, grammar, and readability
- Writing budget narratives
- Creating logic models
AI cannot replace:
- Your program knowledge and impact data
- Relationships with program officers
- Verified facts and statistics (always check)
- Understanding of organizational capacity
- Authenticity — funders read thousands of AI-sounding proposals
Phase 1: Grant Research
Finding Relevant Funders
Prompt for Claude/Perplexity:
"I'm searching for grant funding for [program description]. Our organization:
- Type: [nonprofit/university/small business]
- Location: [city, state]
- Budget: [annual budget]
- Population served: [who you serve]
- Program focus: [health/education/arts/environment/etc.]
- Typical grant request: $[amount range]
Identify types of funders that match this profile. Include:
- Foundation types (community, private, corporate)
- Federal programs to investigate
- State grant programs
- What kinds of information I should look for in each funder's guidelines
I'll use this as a research framework, not a definitive list."Analyzing Funder Guidelines
Prompt (paste the RFP or grant guidelines):
"I need to apply for this grant. Analyze the guidelines and provide:
1. Eligibility check (based on what I share about us)
2. Required sections and page limits
3. Evaluation criteria with weighting if provided
4. Key funder priorities and values (what they care about)
5. Red flags or requirements we might struggle to meet
6. Questions to clarify with the program officer
7. A proposal outline matching their structure
Our organization: [brief description]
Our proposed project: [brief description]
[PASTE GUIDELINES]"Phase 2: Proposal Sections
Needs Statement
The needs statement establishes why the problem matters. Provide your data; AI helps structure it:
Prompt:
"Write a needs statement for a grant proposal.
Organization: [name and mission]
Program: [what we're proposing]
Problem: [the problem we're addressing]
Target population: [who is affected]
Data I have:
- [Local statistic 1]
- [Research finding 2]
- [Community indicator 3]
Funder priorities: [what this funder cares about per their guidelines]
Word limit: [X words]
The needs statement should:
- Open with a compelling hook about the human impact
- Present quantitative data showing the scope
- Connect to local context, not just national statistics
- Show we understand root causes, not just symptoms
- Create urgency without being manipulative
- Transition naturally to our proposed solution"Project Narrative
Prompt:
"Write the project narrative for a grant proposal.
Program description: [detailed description]
Target population: [who, how many, demographics]
Activities: [bullet list of what you'll do]
Timeline: [12-month or multi-year plan]
Partners: [any collaborating organizations]
Staff: [who will implement]
Evidence base: [research supporting your approach]
Funder's evaluation criteria: [from their guidelines]
Word limit: [X words]
Structure the narrative to address each evaluation criterion explicitly.
Use clear headings that signal to reviewers where their criteria are addressed."Evaluation Plan
Prompt:
"Write an evaluation plan for this grant proposal.
Program: [brief description]
Goals: [3-5 program goals]
For each goal, I need:
- 2-3 measurable outcomes
- Data collection methods (survey, database, observation)
- Data sources and timing
- Responsible staff member
- How we'll use data for program improvement
Also include:
- Logic model format (inputs → activities → outputs → outcomes)
- External evaluator mention if required
- How we'll report to the funder
Keep it realistic for a small nonprofit with limited evaluation capacity."Budget Narrative
Prompt:
"Write a budget narrative for a grant request.
Grant request: $[amount]
Total project budget: $[amount]
Other funding sources: [list other funders/amounts]
Budget line items:
- Personnel: [list with names/titles, FTE percentages, salaries]
- Fringe benefits: [rate]
- Consultants: [names, rates, hours]
- Supplies: [list with amounts]
- Travel: [destinations, purpose, estimated costs]
- Other direct costs: [list]
- Indirect/overhead: [rate or amount]
Write narrative justification for each line item. Explain why each expense
is necessary, reasonable, and allowable. Be specific about how amounts
were calculated (e.g., '0.25 FTE × $60,000 salary = $15,000')."Phase 3: Editing and Refinement
Reviewer Perspective Check
Prompt:
"You are a grant reviewer for [funder]. You've read thousands of proposals.
Read this proposal draft and provide:
1. First impression (first paragraph reaction)
2. Clarity score (1-10) with explanation
3. Sections where my argument is weakest
4. Where I've told instead of shown
5. Jargon or unclear language to simplify
6. Missing information a reviewer would expect
7. Whether it reads as written by practitioners who know the work
[PASTE DRAFT]"Specific Section Edits
Prompt:
"Improve this needs statement. Goals:
- Make it more compelling without exaggerating
- Strengthen the connection between data and our solution
- Reduce length by 20% without losing key points
- Ensure it doesn't sound AI-generated (use varied sentence structure)
[PASTE SECTION]"Phase 4: Common Grant Documents
Letter of Intent Template
Prompt:
"Write a letter of intent (LOI) for [grant program].
Our organization: [description]
Proposed project: [description]
Requested amount: $[X]
LOI requirements: [from guidelines]
The LOI should:
- Open with a compelling program description (not organizational history)
- State alignment with funder priorities
- Preview outcomes, not activities
- Close with a clear ask
- Stay under [X] pages"Logic Model
Prompt:
"Create a logic model for this program in a simple table format.
Program: [description]
Population: [who you serve]
Core activities: [list]
Resources available: [staff, partners, budget]
Expected outcomes: [short and long term]
Format: Four columns — Inputs | Activities | Outputs | Outcomes
Include both short-term (1 year) and long-term (3-5 year) outcomes."Grant Writing Best Practices with AI
-
Start with your data. AI structures and writes; your numbers and stories are the core content. Never ask AI to invent statistics.
-
One section at a time. Don’t generate an entire proposal at once. Work section by section, refining as you go.
-
Match funder voice. Read 2-3 grants this funder has funded before writing. Feed their language back into your prompts.
-
Human review every draft. AI proposals often sound similar. Add organizational voice, specific examples, and program staff insights.
-
Verify all facts. Claude and GPT-4 can hallucinate statistics. Every cited number needs verification from its original source.
-
Program officer relationship. AI can’t build the relationship with the program officer that often determines close funding decisions. That’s your job.