How Happy Grants Uses AI
Last updated: September 24, 2026
The short version
Happy Grants uses AI to find, score, and draft. You review, edit, and decide. AI output in Happy Grants is a starting point, not a decision: nothing is ever submitted to a funder on your behalf, and every answer stays editable until you export it.
What AI does in the product
- Grant matching: scores how well each grant fits your organization, from 0 to 100, and explains why.
- Application drafting: writes first-draft answers to funder questions using your organization profile, content library, and uploaded documents.
- Refine and chat: revises answers on your instruction and answers questions about your application.
- Template reading: extracts questions, word limits, and sections from funder templates you upload.
- Profile extraction: builds your organization profile from documents you upload, such as annual reports and past proposals.
- Compliance check and reviewer simulation: reviews your draft against the funder's requirements and scoring rubric.
- Specialized drafting tools: SR&ED narratives, letters of intent, and budget suggestions.
- Grant database upkeep: our crawlers use AI to extract grant details from funder websites, and to review what they extract. See the note on the catalogue below.
How the grant catalogue is built
Catalogue entries are extracted by AI from funder websites and public feeds, not typed in by hand. An entry publishes automatically when the extractor's confidence is high. Lower-confidence entries go to a review queue, where a second AI pass publishes the clear ones and holds anything doubtful for a person to check. So not every grant has been read by a human before it appears here.
To keep that honest: every grant page shows the date we last verified it and links to the funder's own program page. Confirm amounts, deadlines, and eligibility there before you apply, because the funder's page is the authority and ours is a copy. Removing or rejecting a grant is always a human decision. If you find an entry that is wrong or out of date, tell us at privacy@happygrants.com and we will correct it.
Whose AI, and where your data goes
AI features are powered by Google's Gemini models, with Anthropic's Claude models as a backup, called from our servers only. Your API traffic is never exposed to the browser, and neither provider uses your content to train its models.
When you use an AI feature, the relevant content is sent to Google, or to Anthropic if Google is unavailable, for processing in the United States and other countries where they operate. Depending on the feature, that can include your organization profile (mission, sectors, location, size, revenue), your content library, text from documents you upload, the grant questions you are answering, and your chat messages. Your data at rest stays in our Canadian database (Supabase, Canada region). We do not send your payment details or password to any AI provider.
Your role, and your funder's rules
AI drafts can contain inaccuracies or miss what makes your organization distinctive. You are responsible for reviewing, editing, and approving every answer before it reaches a funder. Happy Grants is built around that review step, not as a way to skip it.
Some funders have their own policies on AI-assisted applications, and some ask you to disclose AI use. Check your funder's guidelines before submitting, and disclose your use of Happy Grants where they ask. Honesty wins more grants than polish.
Built the same way
Happy Grants itself was built the way we recommend you work: designed, directed, and tested by a human, with AI assistance throughout development. We use these tools daily, which is exactly why we are clear about what they are good at and where they need your judgment.
More detail
Data handling specifics are in our Privacy Policy, and your responsibilities are in our Terms of Service. Questions? Contact privacy@happygrants.com.
See how it scores your organization
Free plan, no credit card. You review every word.
Get matched free