Industry Guide: Nonprofits

How to Use Claude in Nonprofits: A Guide for Grantmaking, Fundraising, and Program Teams

Stretch a small team further, without cutting corners on donor trust.

This guide is for grant managers, program and development staff, executive directors, volunteer coordinators, and nonprofit operations teams who want to put Anthropic's Claude to work on the proposals, donor communications, and program reports that move mission-critical work forward. It covers Anthropic's own nonprofit-sector releases and what named nonprofit organizations said about using Claude in Anthropic's announcements and customer pages.

Start Learning Claude FreeSet Up Your Practitioner Workspace

Knowing how to use Claude in nonprofits means focusing it on document-heavy, donor-sensitive work: drafting grant proposals and reports from your own program materials, researching funders and matching their priorities to your programs, writing donor and volunteer communications in your organization's voice, summarizing program outcomes for a board or funder, and building internal tools such as intake assistants grounded only in your own reviewed content. Claude drafts, researches, and summarizes. Program staff and leadership own every claim about impact, every dollar figure, and what actually goes out to a funder, donor, or the public, and organizations keep donor and beneficiary data inside their existing privacy and data-handling rules.

Published

Free lesson · no account neededMeet ClaudeWhy Claude?6 minStart the free lesson

Key Takeaways

  • The strongest nonprofit uses are grant and funder research, proposal and report drafting, donor communications, program-evaluation summaries, and volunteer or board documentation.
  • Anthropic runs a dedicated Claude for Nonprofits program with discounted Team and Enterprise pricing, nonprofit-specific connectors, and a free training course.
  • Anthropic's customer pages describe how Brian Impact Foundation, the Epilepsy Foundation, and IDinsight use Claude in their own platforms and workflows.
  • Every AI-drafted funder-facing or donor-facing document needs a human reviewer, and every impact figure needs to trace back to your own program data.
  • Decisions about program strategy, what a grant report claims, and how donor or beneficiary data is handled stay with nonprofit staff and leadership.

Copy and Paste

Prompts You Can Use Today

Grant proposal draft from program materials

Draft a grant proposal narrative for the funder priorities and program description pasted below, using only the facts in our program summary and last year's outcomes report. Match the funder's stated priorities point by point, cite which part of our materials supports each claim, and flag anywhere the funder's requirements aren't addressed by what I gave you. Funder priorities, program description, and outcomes report: [paste].

Grounding every sentence in your own program materials is what makes the draft safe for a program officer to tighten rather than rewrite.

Donor thank-you and stewardship letter

Write a thank-you letter to a donor using only the gift details, the specific program their gift supports, and our organization's voice guide pasted below. Reference the exact program and impact described in our materials, do not invent a dollar figure or outcome beyond what is given, and keep the tone warm but not overstated. Gift details, program description, and voice guide: [paste].

Keeping the claims tied to real gift and program details is what lets a small development team personalize letters at volume without overpromising.

Program evaluation summary for a board meeting

Summarize the program evaluation data below for a board meeting: the outcomes reported, how they compare to last year's figures, any limitation the evaluator noted, and one open question the board should discuss. Cite the specific data point for every claim and do not round numbers in a way that changes their meaning. Evaluation data and last year's report: [paste].

A board needs the real numbers and the real caveats, not a polished summary that drops the uncertainty.

Volunteer onboarding FAQ from your handbook

Build a volunteer onboarding FAQ using only the volunteer handbook and program-specific guidelines pasted below. Write each answer so a new volunteer can act on it without calling the office, reference the handbook section for each answer, and list any question a new volunteer would likely ask that the handbook does not currently answer. Volunteer handbook and program guidelines: [paste].

Flagging the gaps in the handbook is often more useful to a small staff than the FAQ draft itself.

Board meeting minutes and action items

Turn the meeting notes below into formal board minutes: attendees, each motion and its vote, a plain-language summary of each discussion topic, and a clearly separated list of action items with the person responsible for each. Do not add a decision or commitment that is not in the notes. Meeting notes: [paste].

Separating action items from discussion is what makes the minutes usable at the next meeting instead of just a record.

From Anthropic

What Anthropic offers for Nonprofits

Claude for Nonprofits

Anthropic's launch announcement, published in partnership with GivingTuesday, describes a discount of up to 75% on Team and Enterprise plans, access to Claude Opus 4.6, Sonnet 4.5, and Haiku 4.5, new connectors to Benevity, Blackbaud, and Candid, and a free AI Fluency for Nonprofits course through Anthropic Academy covering topics such as grant writing and program evaluation. Anthropic piloted the program with Constellation Fund, Robin Hood, and Tipping Point Community across more than 60 grantee organizations before launch.

· Source

Expanding Claude for Nonprofits

Anthropic's October 2026 update describes a Claude for Nonprofits plugin built with about 30 nonprofits in a design partner circle, installed automatically at no added cost on Team and Enterprise plans, plus a roughly 90% discount for verified nonprofits in low- and middle-income countries, a lower two-seat Team plan minimum, and Claude Code included in nonprofit plans at no added cost. The article says more than 25,000 nonprofits now use Claude, with more than 75% running fewer than 10 seats.

· Source

Claude Corps

Anthropic's program page describes Claude Corps, a fully funded, 12-month fellowship placing up to 1,000 early-career fellows across three cohorts inside at least 400 United States 501(c)(3) nonprofits. CodePath employs and trains the fellows, and Social Finance administers the philanthropic capital and leads measurement. Each host organization must already be a Claude for Nonprofits Team or Enterprise customer, and receives a one-time $10,000 Host Implementation Grant to cover hosting costs.

Source

In Practice

How Nonprofits Teams Use Claude

According to Anthropic's published customer story

Brian Impact Foundation

Brian Impact Foundation, a Korean foundation that awards grants to nonprofits and runs fellowships for individual innovators, built a platform called BEES (Benevolence Enhancing Expert System) on Claude. BEES gathers open-web data on prospective fellowship candidates, uses Claude to extract and summarize the information and assess its quality, and organizes qualifying candidates against the foundation's own selection criteria, with reviewers working through the results in a custom web app. The foundation reports reviewing more than 20,000 candidates last year, about 100 times more than the roughly 200 it reviewed the year before, and tripled the number of reviews per candidate.

Read Anthropic's customer story

According to Anthropic's published customer story

Epilepsy Foundation

The Epilepsy Foundation built Sage, a conversational AI epilepsy assistant that runs as an anonymous widget on its website, currently using Claude Haiku 4.5 on Amazon Bedrock. Sage answers only from a knowledge base of 3,200 pages of content approved by the Foundation's editorial board of epileptologists and neurologists, and refers people to the Foundation's Helpline or partner organizations, such as the Rare Epilepsy Network, when its content does not contain an answer. In under a year, Sage logged about 60,000 interactions, growing to roughly 1,000 unique sessions and 2,700 interactions a week, with average interactions per session rising from 1.2 at launch to 3.

Read Anthropic's customer story

According to Anthropic's published customer story

IDinsight

IDinsight, a research organization that supports global development leaders with data and evidence, reports working up to 16 times faster using Claude, according to Anthropic's Claude for Nonprofits announcement.

Read Anthropic's customer story

According to Anthropic's published customer story

Operation Fistula

Anthropic's nonprofit solutions page describes Operation Fistula, an organization working on obstetric fistula treatment and prevention, using Claude to deliver in about half an hour work that previously took an analyst ten days.

Read Anthropic's customer story

According to Anthropic's published customer story

Mid-Ohio Food Collective

Anthropic's nonprofit solutions page describes Mid-Ohio Food Collective, a Central Ohio hunger-relief organization, using Claude to help rebuild software used by 2,500 food pantries.

Read Anthropic's customer story

According to Anthropic's published customer story

Allies Against Slavery

Anthropic's nonprofit solutions page describes Allies Against Slavery, an anti-trafficking organization, using Claude to analyze more than 750 state anti-trafficking policies.

Read Anthropic's customer story

Illustrative walkthrough

A funder research and proposal sprint before a deadline

  1. A development associate loads the funder's published priorities, the program description, and last year's outcomes summary into a shared workspace.
  2. Claude matches the program description against the funder's stated priorities point by point and flags any requirement the current materials do not yet address.
  3. The associate fills the gaps with the program team and asks Claude to draft the proposal narrative strictly from the completed materials, with a citation to the source for each claim.
  4. The grants manager checks every figure against the outcomes summary and tightens the language to match the organization's voice guide.
  5. The executive director reviews the final draft before it goes to the funder through the organization's normal submission process.

Core Workflows

Claude in Nonprofits Use Cases From Fundraising to Reporting

These are the areas where AI for nonprofits saves the most staff time without weakening the accuracy a funder or donor expects.

Funder and grant research

AI for grant writing starts before the draft: matching a funder's stated priorities to your programs and flagging where your materials do not yet answer what the funder asks.

Proposal and report drafting

Drafting grant proposals and funder reports from your own program summaries and outcomes data, with every claim traceable back to a source document.

Donor communications

Thank-you letters, stewardship updates, and campaign copy written in your organization's voice, grounded in the actual gift and program it describes.

Program evaluation summaries

Turning raw program data into a summary a board or funder can read quickly, without smoothing over the limitations an evaluator noted.

Volunteer and board documentation

Onboarding FAQs, meeting minutes, and policy answers built from your own handbook, so staff spend less time answering the same question twice.

Internal knowledge assistants

Chat-style assistants grounded only in an organization's own reviewed content, built the way the Epilepsy Foundation built Sage on Claude.

What Nonprofits Report

How Nonprofit Organizations Use AI Today

How is AI used in nonprofits? Anthropic's customer pages describe specific organizations building on Claude. Brian Impact Foundation, a Korean foundation that awards grants and fellowships, built a platform called BEES on Claude to gather open-web data on prospective fellowship candidates, extract and summarize it, and organize qualifying candidates against the foundation's selection criteria; the foundation reports reviewing more than 20,000 candidates last year, about 100 times more than the roughly 200 it reviewed the year before, and tripled the number of reviews per candidate.

The Epilepsy Foundation built Sage, a conversational AI assistant that runs as a widget on its website using Claude Haiku 4.5, answering only from a knowledge base of 3,200 pages of content approved by the Foundation's editorial board of epileptologists and neurologists, and referring people to its Helpline or partner organizations when its content does not have an answer. In under a year, Sage logged about 60,000 interactions, growing to roughly 1,000 unique sessions and 2,700 interactions a week. IDinsight, a research organization that supports global development leaders, reports working up to 16 times faster using Claude, according to Anthropic's own announcement.

Across AI for nonprofits, AI for grant writing, AI for donor communications, and generative AI for nonprofits more broadly, the pattern on Anthropic's pages is consistent: AI in the nonprofit sector tends to start with document-heavy, repetitive work, with grant research, drafting, AI for program evaluation, and AI for fundraising freeing up staff time, while program leadership keeps reviewing every figure and claim that reaches a funder, donor, or the public.

Anthropic's Program

Claude for Nonprofits and What It Offers

Anthropic launched Claude for Nonprofits in partnership with GivingTuesday, offering nonprofits a discount of up to 75% on Team and Enterprise plans, access to Claude Opus 4.6, Sonnet 4.5, and Haiku 4.5, and new connectors to Benevity, Blackbaud, and Candid. The launch included a free course, AI Fluency for Nonprofits, taught through Anthropic Academy and built for people with no technical background, covering topics such as grant writing and program evaluation. Anthropic piloted the program with Constellation Fund, Robin Hood, and Tipping Point Community across more than 60 grantee organizations before the public launch.

In October 2026, Anthropic expanded the program with a Claude for Nonprofits plugin built together with about 30 nonprofits in a design partner circle, installed automatically for organizations on Team and Enterprise plans at no added cost. The plugin groups its skills into three areas: extending a small team's capacity for marketing, hiring, and board work; setting up Claude responsibly with organizational mapping and AI-policy guidance; and funding the mission through funder research, proposal drafting, and donor-relationship work. The same update added a discount of roughly 90% for verified nonprofits in low- and middle-income countries, lowered the Team plan minimum to two seats, and included Claude Code in nonprofit plans at no added cost. Anthropic says more than 25,000 nonprofits now use Claude, and more than 75% of them run fewer than 10 seats.

Anthropic has also committed $150 million to Claude Corps, a fully funded, 12-month fellowship that places early-career fellows inside mission-driven nonprofits, run with CodePath as the fellows' employer and Social Finance administering the funding and measurement. The program aims to place 1,000 fellows across three cohorts in at least 400 United States nonprofits, and each host organization, which must already be a Claude for Nonprofits Team or Enterprise customer, receives a one-time $10,000 Host Implementation Grant to cover hosting costs.

Writing in Practice

AI for Grant Writing Done Well

The pattern that holds up in grant writing is a named funder and hard references. Give Claude the funder's own priorities, your program description, and your most recent outcomes data, and ask it to match your programs to the funder's stated priorities point by point rather than write a generic proposal. A draft built that way can be checked claim by claim against your own materials, which is what a program officer or grants manager actually needs before submission.

Donor communications benefit from the same discipline. A thank-you letter or stewardship update should reference the exact gift and the exact program it supports, never a rounded-up impact figure the data does not support. Claude can personalize letters at volume once it has the real gift details and program description for each donor, which is the difference between a scalable process and a form letter.

Keep the division of labor explicit. Claude drafts, researches, and summarizes. The program or development staff member owns the accuracy of every figure and claim, and the executive director or board owns what ultimately goes to a funder, major donor, or the public. Writing that division into your workflow makes AI assistance easier to explain if a funder or donor ever asks how a report was produced.

Choosing Tools

Claude vs ChatGPT for Nonprofit Work

The Claude vs ChatGPT for nonprofits comparison is most useful run on your own documents: a grant proposal that was actually funded, a donor letter that already went out, and a program report a board already reviewed. Compare accuracy against the known-correct version, how each tool handles a gap in the source material instead of guessing, and how closely the draft matches your organization's voice.

For a resource-constrained nonprofit, the decision often comes down to cost and governance as much as output quality: whether a nonprofit discount is available, which connectors reach the donor and grants systems you already use, and whether your board or a major funder has questions about how donor and beneficiary data is handled.

Is it safe to use Claude with donor and beneficiary data? Anthropic's nonprofit solutions page states that nonprofit data is not used to train its models by default, and that Team and Enterprise plans can be configured to meet SOC 2 Type II, ISO 27001, ISO 42001, and CSA Star requirements, with single sign-on on Team plans and a Compliance API and custom data retention on Enterprise plans. Whether a specific setup meets your own board's or funder's data requirements is still a question for your organization to confirm.

Getting Started

A Safe First Month for a Nonprofit Team

Start with a funded proposal

Test prompts on a grant that already succeeded, and compare the AI draft against the proposal that actually went out and was funded.

Load your voice guide and program facts

Put your organization's voice guide, program descriptions, and last year's outcomes data into a workspace so drafts stay consistent and accurate.

Pick one recurring document

Choose one high-volume task, such as donor thank-you letters or funder research, and measure the real change in staff time.

Agree on the data boundary

Decide with leadership which donor and beneficiary data can go into which approved tool, and check your nonprofit discount and data-handling terms before you scale up.

Limits

Where Nonprofit Staff Must Decide

Claude can misstate a figure, smooth over a limitation an evaluator flagged, or draft a claim your program data does not actually support. Every AI-assisted grant report, donor letter, or public statement needs a human reviewer, and every number that reaches a funder or the public should be checked against your own source data first.

Does using AI risk donor trust? The risk is not the drafting tool itself; it is a claim that outruns what your program actually achieved, or donor and beneficiary data going to a tool your organization has not approved. Program strategy, what a report claims, and what gets sent to a funder or donor stay with program staff and leadership.

Treat donor records and any information about the people a program serves as sensitive by default, especially for programs working with vulnerable populations. Use only an approved deployment, follow your own data-handling policy, and route any question about tax status, grant compliance, or legal risk to your organization's own counsel rather than to a draft.

Build Reliable AI Habits for Nonprofit Work

Getting trustworthy drafts from AI is a practiced skill: grounding, source discipline, and review. The free Nightschool AI curriculum is hands-on from the first lesson.

Start Learning Claude FreeSet Up Your Practitioner Workspace

Frequently Asked Questions

How is AI used in nonprofits?

Mostly for document-heavy, repetitive work: funder and grant research, proposal and report drafting, donor communications, program-evaluation summaries, and volunteer or board documentation. Anthropic's customer pages describe Brian Impact Foundation, the Epilepsy Foundation, and IDinsight building on Claude for exactly this kind of work. Program staff keep reviewing every figure and claim that goes to a funder, donor, or the public.

Can AI help write grant proposals?

Yes, for matching your programs to a funder's stated priorities and drafting a narrative from your own program materials and outcomes data. It works best when every claim in the draft traces back to a document you actually gave it. A grants manager still has to confirm the draft matches what the funder asked for and what your program can actually deliver.

Is it safe to use Claude with donor and beneficiary data?

Anthropic's nonprofit solutions page says nonprofit data is not used to train its models by default, and that Team and Enterprise plans can be configured to meet SOC 2 Type II, ISO 27001, ISO 42001, and CSA Star requirements. Whether a specific setup satisfies your own board's or funder's requirements, especially for data about vulnerable populations, is still a question for your organization to confirm.

How much time can AI save a nonprofit team?

It depends heavily on the document type and how complete your source material is, so measure it on your own work. Brian Impact Foundation reports reviewing about 100 times more fellowship candidates after building its platform on Claude, and IDinsight reports working up to 16 times faster, according to Anthropic. Review and sign-off time does not disappear, and it should not.

What discounts does Anthropic offer nonprofits?

Anthropic's Claude for Nonprofits program offers a discount of up to 75% on Team and Enterprise plans, with a deeper discount of roughly 90% for verified nonprofits in low- and middle-income countries. The program also includes nonprofit-specific connectors, a free training course, and, for organizations on Team or Enterprise plans, a nonprofit plugin at no added cost. Check claude.com's nonprofit solutions page for current terms.

What are the biggest AI use cases in the nonprofit sector?

The most common are funder and grant research, proposal and report drafting, donor communications, program-evaluation summaries, and volunteer or board documentation such as onboarding FAQs and meeting minutes. Each works best with a named source document, such as a funder's guidelines or your own outcomes data, and a human reviewer checking the output before it reaches a funder or donor.

Can AI build tools for the people a nonprofit serves?

Yes, when it is grounded only in an organization's own reviewed content. The Epilepsy Foundation built Sage, an AI assistant on Claude that answers only from 3,200 pages of content approved by its medical editorial board, and refers people elsewhere when its content lacks an answer. That grounding discipline is what makes a public-facing assistant safe to run with a small team.

How does Claude compare to ChatGPT for nonprofit work?

Run the comparison on your own finished documents: a funded proposal, a sent donor letter, a board-reviewed program report. Compare accuracy, how each tool handles a gap in your source material, and how closely the draft matches your organization's voice. For most nonprofits, cost, available discounts, and data-handling terms end up mattering as much as a single benchmark.

Keep going

Related Guides and Courses

Related reading

Claude in Nonprofits, by role

Claude for Small Business OwnersClaude for WritersClaude for Operations

Draft Faster, Keep Donor Trust Intact

Hands-on lessons for nonprofit and mission-driven teams.

Browse the Curriculum

Nightschool AI is an independent learning platform and is not affiliated with, endorsed by, or sponsored by Anthropic. Claude is a trademark of Anthropic, PBC. Looking for Anthropic's official Claude Academy? It's at academy.claude.com.