Grant season used to start the same way every year: a blank document, a vague memory of what we said last time, and an afternoon lost to digging through old folders for numbers I knew existed somewhere.
This year I sat down and assembled a complete $100,000 application — narrative, budget, financial summary, every attachment — in a single working session. Not because I'm faster at writing. Because I finally treated grant management the way I treat every other operation: as a system, not a heroic effort.
The tools were Claude and Obsidian. The unlock wasn't the AI. It was the structure underneath it.
Grant writing isn't a writing problem
Here's the thing nobody tells you about grants: the writing is the easy 20%. The hard part is everything around it. Pulling the actual attendance numbers instead of guessing. Finding last year's narrative. Remembering which fund the request comes from and what the funder actually scores. Reconciling the budget against the financials so a reviewer doesn't catch a contradiction. Getting the submission format right so you're not retyping at 11pm on deadline day.
That's not a writing problem. It's an operations problem. And operations problems don't get solved by working harder next year. They get solved by building the structure once.
The setup: Claude + Obsidian
Obsidian is where the organization's memory lives — meeting notes, financial reports, past applications, board input, the whole substrate, all in plain markdown files that sync across my devices. Nothing fancy. Just text, organized.
Claude is the operator that reads across all of it. Instead of me hunting for the ticketing report, Claude pulls the real numbers. Instead of me half-remembering the funder's criteria, it reads the actual policy and the scoring rubric and writes to them. Instead of me eyeballing whether the budget matches the financials, it reconciles them line by line.
The combination matters. Obsidian without Claude is a well-organized pile you still have to dig through. Claude without a structured knowledge base is a confident writer making things up. Together, they're an analyst who has already read everything and doesn't get tired.
Pulling real numbers instead of guessing
The moment it clicked was the budget. In past years, our revenue projections were educated guesses dressed up as a spreadsheet. This time, the numbers came straight out of the systems that already track them — actual ticket sales across 81 performances, real festival attendance, the prior year's filed financials.
A funder can feel the difference between a budget built on hope and a budget built on history. One reads as a wish. The other reads as an organization that knows exactly what it does.
It also caught my own mistakes. At one point the draft implied the city had funded a building project first and our community matched it. Claude flagged that the public record showed the opposite — we raised the money first, then the city matched. That's not a small edit. Getting the sequence backwards in front of the people who were in the room would have cost credibility. The correct version was both more accurate and a stronger story.
That's the quiet value of a second set of eyes that has actually read the source material: it protects you from the errors you can't see because you're too close.
The unlock: a grant playbook that compounds
Here's the part that changes everything, and it has nothing to do with this year's application.
When we finished, I had Claude write a grant playbook — one markdown file that captures everything durable about this specific grant. The submission mechanics. The field-by-field mapping of what goes where. The scoring rubric and the single thing the funder cares about most. Which scripts pull which numbers. The gotchas we hit, including a legal-address typo that's been quietly copied forward for years.
Next year, I don't start from a blank page. I start from the playbook. The thinking is already done. The structure gets copied, the numbers get refreshed, and an afternoon of archaeology becomes twenty minutes of updates.
This is the principle I build everything on: build it once, use it everywhere. A one-time application is a cost. A repeatable system is an asset. The difference between them is whether you capture what you learned or let it evaporate the moment you hit submit.
Most organizations let it evaporate. The knowledge lives in one person's head, and when that person is busy — or gone — the next grant cycle starts from scratch again. A playbook turns institutional memory into a file anyone can pick up.
Why this matters for small nonprofits
Big institutions have grant departments. The 200-seat arts center, the local shelter, the community foundation with three staff — they have one person who already does four jobs, writing grants in the cracks between everything else.
That's exactly the organization that can't afford to rebuild knowledge every year, and exactly the one that usually does. The fix isn't a bigger team or expensive software. It's a structure that makes the work repeatable and a tool that can hold the whole context at once.
Order is the foundation of freedom. Get the structure right — your numbers where you can find them, your funders' requirements written down, your hard-won lessons saved in a playbook — and the work that used to eat your spring becomes a Tuesday afternoon. You're not freed from the grant. You're freed from doing it cold, every single time.
I'm Heath Johnson. I run a 120-year-old opera house in Spearfish, South Dakota, and through my consulting practice, Groundwork, I help mission-driven organizations design the operational systems that let small teams do the work of large ones. If grant season — or any season — keeps eating more of your time than it should, that's usually a structure problem, not a workload problem. And structure problems are fixable.