- Where it gets stuck
- Each client has its own eligibility: a registered nonprofit, a research lab with an institutional partner, a company under a size limit. New calls appear on Grants.gov, foundation sites and newsletters every week, and matching them against every client depends on whoever has time to read them, so a good fit is often found three weeks before the deadline.
- What the workflow does
- Each new call from the sources you follow is read against a profile per client: eligibility, geography, match funding, typical award size and past results with that funder. A short fit note per client gives the funder, the deadline, the eligibility points to confirm and the reasons the call may or may not fit, each linked to the call text. The consultant decides which calls to raise with the client, and the message to the client is drafted for the consultant to approve.
AI for grant writing
Three funder deadlines land in the same week, and the budget justification for the biggest one is still in last year’s file.
We draft from each client’s boilerplate library and past applications, and the consultant approves every section before it reaches the client.
AI grant writing works best on the three jobs around the argument in a grant consultancy: watching funding calls against each client’s eligibility, drafting the standard sections of an application from the client’s own boilerplate and past submissions, and keeping post-award reports and spending evidence on time. The workflow starts from the documents and tools you already use, nothing reaches a client or a funder without the consultant’s approval, no client file is pasted into a public chat tool, and the case for funding, the final wording and the submission stay with your consultants and the client who signs.
Today
- Three deadlines this weekOrganisation history, team bios and budget justification rebuilt from last year’s filesSenior writer’s weekend
- New foundation callSpotted in a newsletter, two clients may be eligibleFound three weeks before the deadline
- Interim report due FridaySpending evidence and outcomes data still with the clientFourth reminder
With the workflow
- Fit note per clientFunder, deadline, eligibility points to confirm and the client’s past results with that funderConsultant picks the calls to raise
- Draft sectionsOrganisation history, team, budget justification and attachments checklist in the funder’s template, each line linked to its sourceWriter edits before the client sees it
- Report calendarEvery report date from the grant agreement, with the evidence request drafted ahead of timeRequest sent after approval
Where can AI save a grant writer hours on each application?
Each job comes back with every call, every application or every award, relies on documents the consultancy already holds and ends with a decision by the consultant.
Funding calls are found late, or for the wrong client
Every application starts again from a blank template
- Where it gets stuck
- Organisation history, team bios, past results, the theory of change, the data management plan and the budget justification are rewritten for each funder’s format and word limits, from files scattered across past applications. The senior writer spends the week on sections the client has already approved three times, and the argument that wins the award gets the last evening.
- What the workflow does
- A first draft of each standard section is prepared in the funder’s template and within its word limits, from the client’s boilerplate library and past applications, with each sentence linked to the document it came from. The budget justification is drafted line by line from the client’s budget spreadsheet, and an attachments checklist shows what the funder requires and what is still missing. The consultant writes the case for funding, edits every section and decides what goes to the client.
Post-award reports surface the week they are due
- Where it gets stuck
- An award brings interim reports, a final report, spending evidence and sometimes a change to request from the funder, each on its own date in the grant agreement. The dates sit in a spreadsheet, and the program lead at the client hears about the evidence needed when the deadline is close, while the next payment depends on it.
- What the workflow does
- Each award gets a calendar built from the signed grant agreement, with the report dates, the evidence each report needs and who holds it at the client. Ahead of each date, the request for outcomes data and spending evidence is drafted in the consultancy’s tone, and a first draft of the progress report is prepared from what came back and from the objectives in the original application. The consultant approves the requests and edits the report; the client signs and submits it.
Which flow should you start with?
Three questions take about a minute, and the answer points to one of the moments above, or to a smaller step when a workflow would be too much.
Answer the three questions to see where to start.
This is an exercise: nothing is sent when you use this selector.
Do we have to change how we store client files?
No. The workflow reads the funding sources, client folders and spreadsheets your consultancy already uses, prepares a fit note, a section or a request and puts it back where your team works, once the consultant has approved it.
- A call, a deadline or a client email arrivesA new funding call, an application due date, a report date in a grant agreement.
- The client file is readEligibility profile, boilerplate library, past applications, budget and grant agreements, within the access you allow.
- A note, a section or a request is preparedFit note, draft section, budget justification or evidence request, each point linked to its source.
- The consultant decidesYour edits stay inside your consultancy’s own deployment, so the next draft starts closer to what your writers would have produced.
- Back into your toolsThe draft is filed in the client folder and the request leaves from Outlook.
- Instrumentl
- GrantStation
- Grants.gov
- SAM.gov
- Google Drive
- Microsoft Word
- Excel
- Outlook
Three conditions before anything acts
Your existing tools
The workflow runs inside the tools you already pay for, with the access rights they already have, so the team keeps its usual screens.
A named person approves
Every draft waits for a named person. Nothing goes out in the consultancy’s name without that approval, and exceptions go back to a human.
Every step leaves a trace
Sources, drafts, corrections and approvals are recorded, so you can see why an output was prepared and improve the next one.
The path from here: three steps
The free Leverage Scorecard
You fill it in online, without a call. The result shows at once, and only the detailed report asks for a work email.
A 30-minute call
You tell Loïc Guyon, the founder, where the time goes, and together you see whether a workflow is worth testing.
The Leverage Assessment, $999, excluding tax
A review of where you stand today and where AI can help, in two 45-minute sessions with a written report between them. If it is useful, we test one or two workflows on your own examples, anonymised or invented when the data is sensitive. The fee is deducted in full if you continue to a build, with no obligation to build. It is invoiced by AI Jungle LLC, a US company.
Looking for a specific product?
PitstopThe shared workspace for clients, funders, applications and deadlines, so an award’s history is readable without asking the writer who handled it.Discover Pitstop →Managed AI agent serviceWe build, run and maintain one defined workflow, such as the funding watch or post-award reporting, with named approvers, a written scope and handover notes.See the managed service →
Before you decide, your senior writer marks our draft against one of your own past applications
AI outreach case study: Asia-Connect. In 3 months of production use: 755 executives contacted, a 24% reply rate, 139 qualified meetings booked, 90 of them held, and 0 messages sent without the partner’s approval.
“In the last three months, the agent helped me to book more than 90 qualified meetings with executives I would never have had time to reach, and I approved every single message before it went out. It works while I am in meetings.”
During the Assessment we draft the standard sections of one application you have already submitted: you send the funder’s guidelines, the client’s boilerplate and budget spreadsheet as they stood before that application, under a confidentiality agreement signed first, with the client’s name replaced by a code. We return the draft sections, the budget justification and the attachments checklist we would have produced. Your senior writer compares them with what was submitted and marks each section usable, needs work or wrong in about an hour, and you keep that result whether or not you go ahead. Approval works the same way in production: at Asia-Connect, an executive search firm, the partner approved every message before it went out. Read the case study
Questions grant consultants ask before a call
Where do our clients’ applications and budgets go?
No client application, budget or grant agreement is pasted into a public chat tool. The workflow reads only the folders, mailboxes and accounts you name, and any step your consultancy judges incompatible with a client contract or a funder’s confidentiality terms stays manual.
- Your data is hosted in Germany, on Hetzner, and none of it is used to train a model.
- If you have a constraint such as professional secrecy, sensitive data or a contractual requirement, processing stays in the European Union; otherwise we choose, for each task, the model with the best price-to-intelligence ratio, and you get the provider and the place of processing in writing before it reads anything.
- A data processing agreement is signed before anything is set up or sent, including for the workflows tested during the Assessment, and you receive the written list of the systems that will read your data.
- Your installation is kept separate from every other client’s.
- Everything is yours, your data and everything produced for you: at the end of the contract you receive a full export, then all of it is deleted from our systems.
- A person approves everything before it is sent, and you can refuse any step, which then stays manual.
Which grant writing consultancies is this for?
Owner-led grant writing and funding consultancies, and independent grant writers with a small team, who work for nonprofits, research teams or companies and still track calls, assemble standard sections and chase post-award evidence themselves. The workflows cover that work around the application, never the case for funding itself.
Do funders accept applications drafted with AI?
Funders’ rules differ, and some now ask applicants how AI was used, so the guidelines of each call are read and any restriction goes into the fit note. The workflow drafts the standard sections from the client’s own material, the argument and the final wording are written by your consultants, and the client remains the applicant who signs and submits. Where a funder rules out AI-assisted drafting, that application stays manual.
Instrumentl already finds grants for us. What do you add?
Keep it: Instrumentl and GrantStation find and track opportunities well. The workflow adds what comes after a match: the fit note against each client’s eligibility and past results with that funder, the first draft of the standard sections from the client’s own library, and the report and evidence dates of each award, with the requests drafted. When your prospecting tool already holds the pipeline, the fit note starts from it.
What should we start with?
Usually the standard sections of an application that comes back every year: they repeat with every funder, the client’s material already exists and a senior writer sees at once whether a draft is usable. If calls are found too late, start with the funding watch. Use the selector on this page, then bring one past application to a 30-minute call.
Who is responsible if a section or a report is wrong?
Your consultancy and the client who signs, exactly as today, which is why nothing reaches a client or a funder unapproved. Each draft keeps the source of every figure and claim, so a writer can check it before it goes out. The contract sets out in writing what AI Jungle answers for, and we give you wording to check with your professional indemnity insurer and your client contracts.
How much does it cost?
The Leverage Assessment costs $999, excluding tax, and the amount is deducted in full if you continue to a build. It is invoiced by AI Jungle LLC, a US company. The Leverage Scorecard is free, and the full path is described above.
In 30 minutes, we pick the past application we will work on in the Assessment.
If you book the Assessment, you send that application’s guidelines, the client’s boilerplate and its budget spreadsheet, with names replaced by codes, under a confidentiality agreement signed first. By the end of the call you know what to send and when you get the result back.