Claude Project: Grant Writing Assistant with School Data

What This Builds

You'll build a dedicated Claude Project configured as a grant writing assistant for your school, loaded with your school's demographic data, past funded grants, approved narrative language, school outcome history, and the "school story" you want funders to know. Once built, every new grant application starts with a fully contextualized assistant that knows your school's evidence of need, your program approaches, your outcomes data, and the language that has worked with funders before. An 8-hour grant write drops to 2 to 3 hours, and you stop rewriting the same "evidence of need" section from scratch for every grant.

Prerequisites

• A Claude Pro account ($20/month, required for Claude Projects)
• 2 hours to gather grant documents and configure the project
• Comfortable using Claude for grant narrative drafting (Level 3 guide)
• 2 to 3 past successful grant narratives (if you have them, ideal but not required)
• Your school's outcome data: academic results, attendance, demographics, program descriptions

The Concept

A grant writing assistant project is like having a development director on your team who has read every grant you've ever submitted and knows your school's story cold: the demographics, the need, the outcomes, the program design. When a new RFP comes in, you show them the requirements and they draft the narratives based on everything they already know about your school. The only thing you add is the specific data for this grant cycle and the specific requirements of this funder.

Build It Step by Step

Part 1: Gather your grant library materials

Before setting up the project, collect:

1. School data document: Create a one-page Google Doc with:

   • Current year demographics (enrollment, FRL%, ELL%, disability%)
   • Academic results (MAP, state test, 2 to 3 years if available)
   • Attendance and graduation/promotion data
   • Teacher credential and retention data
   • Any awards, recognition, or notable achievements

2. Evidence of need narrative: A 300-word description of why your community needs your school. Include neighborhood poverty data, achievement gaps, and what students' lives look like without your school. This is the section you rewrite for every grant. Save the best version.

3. Program description: A 400-word description of your school's educational approach. What makes your program distinctive? What do you do that works?

4. Past successful grant applications: If you have 1 to 3 funded grants, pull the approved narrative sections. These are your proven language. Funders said yes to these specific words.

5. Grant outcomes data: How have past grants performed? What did you accomplish with the funds? Even rough notes are useful.

Part 2: Set up your Claude Grant Writing Project

1. Log into claude.ai (Claude Pro) → Projects → New Project
2. Name it: "Grant Writing [School Name] [Year]"
3. In the Project instructions section, paste a customized version of this:

You are a grant writing assistant for [School Name], a charter school in [city/state].

YOUR ROLE:
- Help evaluate grant fit before writing begins (save time on bad fits)
- Draft narrative sections based on our school's evidence, programs, and outcomes
- Maintain consistency across all narrative sections of an application
- Match tone and language to what has worked for us with funders before

SCHOOL PROFILE FOR GRANT APPLICATIONS:
[Paste your school data document here — demographics, outcomes, academic data]

OUR SCHOOL STORY (use this framing across all grants):
- Who we serve: [brief description of student population]
- Why our community needs us: [brief evidence of need — poverty rates, achievement gaps]
- What we do: [brief program description — your educational approach]
- What outcomes we achieve: [your strongest results]
- Our values: [mission and values relevant to funders]

GRANT WRITING GUIDELINES:
- Evidence of need: Lead with community data, not just school data — funders want to understand the gap our school addresses
- Program design: Be specific — name the interventions, the dosage, the staff who deliver them
- Budget requests: Always explain the why behind each line item
- Outcomes: Propose measurable outcomes that you can actually track and report
- Tone: Mission-driven but evidence-based — confident without overpromising
- Avoid: Generic language that could describe any school; overly academic jargon; vague claims without data

LANGUAGE TO USE (adapted from our funded grants):
[Paste 3-5 sentences from your best funded grant narratives — the exact language that worked]

LANGUAGE TO AVOID:
[List any phrasings that you've been told don't land well, or that feel inauthentic to your school]

Part 3: Upload your knowledge files

In the Knowledge section, upload:

1. Your 2-3 best funded grant applications (full PDFs or text files)
2. Your school data document
3. Your current program description

Why: When you ask Claude to draft a new "evidence of need" section, it can look at what language worked for you before and build from it rather than starting generic.

Part 4: Test your grant assistant

Start a new conversation inside the project and test it with:

Test 1: Fit assessment: "Here is a grant RFP I'm considering. Based on our school profile, assess our fit on a scale of 1-10 and tell me the top 3 reasons to apply or skip: [paste short RFP summary]"

Test 2: Evidence of need: "Draft a 500-word Evidence of Need section using our school profile. Target funder type: corporate foundation focused on workforce development and college readiness."

Test 3: Program design: "Draft a 400-word Program Design section describing how we would use a $30,000 literacy grant to add a K-2 reading intervention block. Specific plans: 30 minutes daily, taught by a reading specialist we would hire, targeting students 1+ years below grade level."

Evaluate: Does the output draw on your school's actual data and story, or does it sound generic? Refine your project instructions until the outputs sound like they came from someone who knows your school.

Real Example: A Three-Grant Season

Setup: You've built your grant project with your school profile, evidence of need narrative, program description, and 2 funded grant applications as knowledge files.

Grant 1: Literacy Foundation RFP ($20,000) arrives: Open project, new conversation. "Here is the RFP. Does this fit our school? [paste RFP]" Response: Fit score 8/10. We clearly serve the target population, our need data is strong, our proposed use is appropriate. "Great. Draft the Evidence of Need section (600 words) and the Program Design section (500 words)." Result: Two draft narratives in 3 minutes that draw on your school's actual data and reference language from your previous funded grants. Review and refine: 30 minutes. Total writing time for this grant: 2 hours vs. 8 hours.

Grant 2: Community Development Foundation ($50,000, complex) arrives: Same project, new conversation. "Fit check: [paste RFP]" Response: Fit score 5/10. This funder focuses on community development outcomes, not direct academic services. Our school's work doesn't map cleanly to their priorities. Decision: Skip this grant. 0 hours wasted.

Grant 3: State Literacy Grant ($35,000) arrives: New conversation. Fit score 9/10. Narratives drafted: 4 sections in 15 minutes. Total writing time: 2.5 hours.

Season total: 3 grants evaluated in 30 minutes combined. 2 applications submitted in 4.5 hours total writing time vs. 16+ hours without the project.

What to Do When It Breaks

• Narratives sound generic despite the project setup: Your knowledge files may not be fully indexed. Try copying key narrative text directly into the project instructions instead of relying on uploads.
• Claude doesn't reference your past grants: Paste specific approved language directly into the instructions section under "Language to use from our funded grants." Don't rely on the uploaded files for this.
• Output budget narratives are inaccurate: Always specify the exact dollar amount and exactly what you plan to buy. Budget narratives require specific numbers you must provide.
• Project instructions are too long: Claude Pro projects support long instructions, but if your instructions exceed 3,000 words, prioritize the school data and writing guidelines sections over the program description.

Variations

• Simpler version: Build a Google Doc with your school's "grant data sheet" (1 page of key stats and program descriptions) and paste it at the start of every grant chat in regular Claude.
• Extended version: After each grant cycle, add the successful narratives back into your knowledge files. Over time, the project learns which language has worked.

What to Do Next

• This week: Build the project and run your next incoming RFP through the fit assessment
• This month: Draft 2 full grant applications using the project and compare your time investment to past grants
• Advanced: Connect your grant project to your school voice project. Use the school voice project for communications and the grant project for funding applications. They share context but stay focused on their purpose.

Advanced guide for charter school administrator professionals. These techniques use more sophisticated AI features that may require paid subscriptions.