Skip to content

How it works

From your own records to a finished application

Grant Autopilot does the searching, drafting and tracking that a grant writer would. It only says what your own documents and interviews support, it shows you where every fact came from, and your team decides what gets sent.

Step 1

Step 1: Build your fact base

Everything Grant Autopilot says about your organization comes from your fact base: short, specific facts taken from your own materials, each with the words it came from.

  • Upload what you already have. Annual reports, budgets, financial statements, program descriptions, past applications (funded or not) and letters of support. PDFs, Word files, spreadsheets, text and images all work.
  • Or talk it through. A guided interview asks about your mission, programs, the people you serve and your results. Type your answers or speak them and we'll transcribe them. You can also upload notes or a transcript from a conversation.
  • Every fact keeps its source. Each one records the exact quote and the document or interview it came from, so anyone can check it.
  • Your team confirms what's true. Facts start as suggestions. Someone on your team confirms, corrects or rejects each one, and rejected facts are never used.
  • Sensitive details stay out. Mark facts about individual clients or participants as restricted, and they're never sent to the AI when writing applications.
  • See what's missing. A coverage view shows which topics funders commonly ask about are thin or missing, so you know what to add next.
Example facts for a fictional food pantry. Each keeps the words it came from, and the last one is waiting for someone to confirm it.

Step 2

Step 2: Find grants that fit

We look for opportunities in four places, then check each one against your fact base so the most promising ones rise to the top.

  • Federal grant listings

    Every day we check the official Grants.gov listings for new and forecasted federal opportunities, straight from the agencies that post them.

  • Your funder watchlist

    Add the funding pages of foundations, state agencies and community funders you care about. We re-check them on a schedule and let you know when something new is posted.

  • Imports by link or RFP

    Found an opportunity yourself? Paste its link or upload the request for proposals. We read the requirements from it, and imported grants stay private to your organization.

  • Foundation prospecting

    We look through private foundations' public 990-PF tax filings for funders whose past grants resemble your work. These are leads to research and build relationships with, not open calls for proposals.

How we rank what we find

  • Each opportunity gets an eligibility check, a fit score and an estimate of how likely an award is, with the reasons spelled out.
  • Every reason cites the facts it relies on, so you can check our thinking.
  • When eligibility depends on something we don't know, such as an active SAM.gov registration, you get a plain question instead of a guess.
  • We only list what a funder has actually published. We don't guess deadlines, amounts or eligibility rules.

Step 3

Step 3: Review a cited draft

When you decide to apply, Grant Autopilot reads the funder's questions and drafts each section from your facts. Then a separate check reads every claim before a person ever sees it.

  • The drafter works only from your fact base, plus your own past applications for tone and style.
  • Every factual sentence lists the facts it relies on.
  • A separate checker, run as its own step, compares each claim with the cited facts and their quotes, and makes sure every number, date, dollar amount and percentage actually appears there.
  • Sentences that only set the scene can't contain numbers, names or outcomes, so nothing slips through as “just framing”.
  • Anything unsupported becomes a flag for a person. Questions your records can't answer become gaps, and gaps become interview questions for your team.
  • You can edit or regenerate any section, or ask for changes. Each change is saved as a new version and checked again.

A look at the review screen

This is what a reviewer sees once the checker has run. It's a picture, not a live screen, put together from the same parts the app uses.

An example review screen for a fictional food pantry, after the checker has run. The numbers match the notes below.

  1. A citation on every factual sentence

    Each small number points to a fact in your fact base. Open it to see the fact, whether your team has confirmed it, and the exact words from the document it came from.

  2. Unconfirmed facts show in amber

    This fact was found in one of your documents, but nobody on your team has confirmed it yet. The sentence can stay, but it's flagged, and autopilot can never rely on it.

  3. Unsupported sentences are flagged

    The independent checker found nothing in your records behind this sentence, so it's underlined in red and listed as a blocker until someone edits or removes it.

  4. Gaps become questions, not guesses

    The funder asked something your records don't answer. Instead of inventing a number, the draft asks your team. You can answer in a short typed or spoken interview.

  5. A fact-check you can read

    A count of what the checker looked at: supported claims, number mismatches, gaps and citations to unconfirmed facts. It only passes when nothing is unsupported or missing.

  6. The right level of review, and your approval

    Open blockers send this application to full review. “Approve & submit” stays disabled until they're resolved, and nothing is sent until a reviewer approves it.

How much review each application gets

Each application is routed by its value and how likely it is to be funded, using limits your organization sets. Bigger and more promising requests get more of your attention.

  1. Full review

    Large requests, ones you're likely to win, or anything with unresolved flags

    Your team goes through the application section by section, with every flag and source in view.

  2. One-click approval

    Most applications

    A clear summary with any open questions highlighted. Approve when you're satisfied, or open the full draft.

  3. Autopilot (optional)

    Small, low-risk requests that pass every check

    Only if an owner or admin turns it on and sets its limits, such as a maximum amount and a monthly cap. Anything flagged, unconfirmed or high-value still comes to you.

Nothing is sent without your approval. That's the default for every organization. Autopilot stays off until an owner or admin turns it on and sets its limits, and it can be turned off again at any time.

Step 4

Step 4: Approve, submit, follow up

Once a reviewer approves, Grant Autopilot sends or packages the application, keeps an exact record of it and keeps track of what comes next.

  • Sent by email

    Where a funder accepts applications by email, we send the approved application.

  • Packaged for portals and Grants.gov

    For online portals and Grants.gov, we prepare the complete package and a step-by-step guide. Your authorized representative submits it, then marks it done and adds any tracking number. Direct submission to Grants.gov isn't supported yet.

  • An exact record

    We keep the exact package that was approved and submitted, with a digital fingerprint that shows it hasn't changed since.

  • Follow-ups tracked

    Deadline reminders, status checks, report due dates and thank-you notes are tracked, and you can record whether each application was funded.

Honest limits

What Grant Autopilot won't do

It's a tool for your team, not a replacement for your judgment.

  • It won't invent facts, numbers or stories to fill a gap. It asks you instead.
  • It won't send anything without a person's approval, unless your organization turns on autopilot for small requests and sets its limits.
  • It can't submit directly to Grants.gov yet. It prepares the package for your authorized representative.
  • It can't promise funding. Funders make their own decisions, and your organization remains the applicant.