AI Prompt Library

Ready to use AI prompts for common grant, finance, and partnership tasks with guidance on how to use them safely.

The prompts (10)

Program Design

Build a problem tree from assessment notes
Use this AI prompt to turn raw field notes into a structured problem tree with supporting evidence for your project design.

Proposals and Fundraising

Draft a Concept Note from Project Notes
Get an AI prompt to turn rough notes into a structured concept note, designed to reduce invented data, with checks for you to run.

Grants Management and Compliance

Extract Obligations from a Grant Agreement
A copy-paste AI prompt that pulls the responsibilities, deadlines and prior-approval triggers out of a signed funding agreement into one table you can check and track.
Explaining Budget Variances to Funders
A copy-paste AI prompt that builds a variance table and a plain-language narrative for your funder, and flags where a reallocation may need approval.

Finance and Budgeting

AI prompt for grant budget narratives
A copy-paste prompt and hand-check list that turn a line-item budget into clear budget notes with rules against inventing numbers.

Safeguarding and Accountability

Review a safeguarding policy for gaps
A structured method and AI prompt to identify missing elements in your safeguarding or PSEAH policy against sector standards.

Partnerships and Locally Led Work

Partner Due Diligence Questions
Get a ready-to-use AI prompt and a guide to create proportionate due diligence questions and evidence requests for local partners.

Organization: Governance, People and Operations

Terms of Reference for Consultants
A practical guide and AI prompt to help nonprofit managers draft clear, professional terms of reference that attract qualified consultants and improve the chance of a good result.

Humanitarian Response

Draft a situation report from field notes
An AI prompt that turns rough field notes into a concise humanitarian situation report, with a note line cited for every figure and a list of gaps instead of guesses.

AI for Nonprofit and Development Work

Verify AI Drafts for Invented Facts
Learn how to use a verification prompt that helps you check each claim in an AI-generated report against your own source documents.

What This Covers

This library is for program, grants and finance staff at nonprofits. The prompts cover the working tasks of the grant cycle: concept notes, budget narratives, pulling obligations out of an agreement, explaining budget variances, partner due diligence and consultant terms of reference. Others help you review a safeguarding policy, build a problem tree, draft a situation report and check an AI draft against its source. The aim is a decent first draft in place of a blank page, and a habit of checking that draft against the source.

Start Here

Pick the task you are stuck on. Each prompt gives you a structured first draft, not a finished document.

  • Concept note drafting. Turns a project idea into a brief outline for a potential funder, with title, geography, population focus and budget.
  • Budget narrative. Drafts a justification saying why each item is necessary and how its cost was calculated.
  • Obligation extraction. Pulls the responsibilities, deadlines and prior-approval triggers out of a grant agreement into one table you can check, each row with the quote it came from.
  • Variance analysis. Explains why actual spending differs from budget, separating timing variances from genuine problems.
  • Partner due diligence. Builds a checklist for deciding whether a proposed partner can manage funds and deliver safely.
  • Consultant terms of reference. Drafts a consultant's scope of work, deliverables, timeline and qualifications.
  • Safeguarding policy review. Checks a policy you already have for gaps.
  • Problem tree. Turns assessment notes into causes and effects.
  • Situation report. Drafts an update from field notes.
  • Check an AI draft. Helps you spot invented facts in anything an assistant wrote.

To use one, paste the prompt and the relevant source text, then check the output line by line against the source. Match every quoted clause and date to the original text.

Before you paste anything, check your organization's AI use policy and which tools it approves. Keep confidential material, such as an unpublished grant agreement or a case file, out of public AI tools, and use a tool your organization controls. Many grant agreements carry their own confidentiality terms, so read those first.

Key Concepts

  • Concept note: a brief outline of a project idea, written for a potential funder before any full proposal, engaging enough to make them want more.
  • Results chain: the hypothesized link between inputs, activities, outputs, outcomes and impacts (see program design).
  • Budget narrative: the explanation behind the line items, saying how each was calculated and why it is necessary for project goals.
  • Budget variance: the gap between budgeted and actual spending. Favorable means spending came in under budget, unfavorable means over.
  • Obligation extraction: pulling the responsibilities, deadlines and dependencies in an agreement into one place.
  • Due diligence: a structured review, done before signature, payment or legal commitment, of whether a partner can manage funds and deliver safely.

Templates

Organize your material this way before asking an assistant to draft.

  • Concept note overview block: Title | Geography | Population focus | Number and type of people served, direct and indirect | Duration in months | Budget with currency.
  • Safeguarding policy workplan (one source's suggested list): Objectives | Activities | Outputs | Indicators | Responsible personnel | Timeframe | Budget.
  • Partner due diligence checklist: Public source review | File evidence review.
  • Compliance file checklist: administrative, financial and activity-specific records, including the activity and budget description, the eligibility determination, governing policies, signed partner agreements, procurement records and payment requests.

AI for This Topic

An assistant can turn rough notes into a structured first draft. It can draft budget justification text, but whether each cost is reasonable and necessary is your judgment.

For obligation extraction, paste the agreement as text, not scanned images, so every clause can be read. Ask for the clauses themselves, not a summary.

Two rules: keep confidential material out of public tools, and check every output against the source before you use it.