A Funding Finder for Michigan Farmers At Any Experience Level

How my own frustration with the grant system, dozens of conversations with farmers across Michigan, and an AI code partner became the Rural Funding Finder.

At a glance

• What it is: A tool that asks what you’re trying to build on your farm, then finds real funding and tells you honestly what kind of money it is.
• My role: Product vision, research, positioning, and hands-on direction of the build.
• Built with: An AI coding partner, directed and reviewed by me.
• Live at: funding.altocitylimits.com
• Coverage today: Michigan and federal programs, with more states to come.

The search box that asks the wrong question

The first time I tried to help someone find agricultural funding, I did what anyone does. I opened a government website and started reading. An hour later I had twelve browser tabs, four acronyms I couldn’t keep straight, and no idea whether the person I was helping could actually apply to any of it.

The problem wasn’t a lack of information. It was the opposite. Every program had a page. Every page assumed you already knew the program’s name. The whole system is organized around the answer — which program do you want? — when the person standing in the barn only has the question: I want to do this thing. Is there money for it?

That gap stuck with me. I spent years as a journalist learning to translate complicated things into plain language, and then years in product marketing figuring out how to position a product so the right person understands why it matters in about six seconds. Looking at the funding system, I saw the same failure I’ve been paid to fix my whole career: a product that describes itself by how it’s built instead of what it does for you.

So before I wrote a word of a plan, I did the thing that actually matters. I went and listened.

What dozens of farmers actually told me

I talked with farmers across Michigan — beginning farmers, multi-generation operations, a few homesteaders, one dairy operator who’d been through the application wringer more than once. I wasn’t collecting testimonials. I was trying to find the pattern underneath the frustration. A handful of themes came up again and again, almost word for word.

“They call everything a grant.” This was the big one. People would say they were “looking for grants” when what fit their project was a cost-share program, or a low-interest loan, or a reimbursement they’d have to front the money for. The word grant had become shorthand for free money I probably won’t get, and it was quietly setting everyone up for disappointment. Cost-share isn’t a grant. A loan guarantee isn’t a grant. Getting that wrong wastes weeks.

The deadlines that weren’t real. More than one person told me they’d found a “perfect” program online, only to discover the round had closed months ago — or that the date they were counting on was something a blog had guessed at, not something the agency ever published. In 2026, for instance, the well-known federal REAP grant was paused while the rules were updated. Plenty of stale web pages still made it look wide open.

Eligibility surprises at the worst moment. A homeschooling family gets excited about an education-and-agriculture grant, spends a weekend on it, and only then learns that individual families can’t apply directly — it has to run through a nonprofit, a school, or a co-op. That’s not a small footnote. That’s the whole thing, and it should be the first thing you’re told, not the last.

“I don’t have a program. I have a project.” Nobody wakes up wanting to apply to EQIP. They want to fence a pasture and run water for rotational grazing. They want to put up a 30-by-40 barn. The funding system has no idea what to do with a project — it only understands programs. Someone has to do the translation from “here’s what I’m building” to “here are the three things that might pay for pieces of it,” and right now that someone is usually nobody.

None of this was a technology problem. It was a clarity problem. And clarity is a thing you can design.

The idea: ask what you’re trying to do

The product thesis wrote itself out of those conversations. Instead of asking which program?, the tool asks what are you trying to build? — in plain language — and works backward to the money.

A few principles became non-negotiable, and I want to be specific about them because they’re the whole point:

Every result is labeled by its real funding type — grant, cost-share, loan, loan guarantee, reimbursement, tax credit, rebate, scholarship, or technical assistance. When there’s no open grant that fits, the tool says so out loud instead of dressing up a loan as free money.

It breaks a project into fundable pieces. “Fence and water for grazing” comes back as the cost-share program that actually funds that work, not a keyword match on the word “farm.”

It never invents a deadline. If we don’t have an official date from the agency, the program is marked by its real status — open, continuous, expected, closed — and the tool tells you what to do about it. Honesty about when turned out to be as important as honesty about what.

And it leads with eligibility. If individual families can’t apply to something directly, that shows up as a flag on the result, with the note that a nonprofit, co-op, library, or school may be able to apply on the project’s behalf. The disappointing news comes first, while it’s still useful.

Under all of it: the tool never makes up a fact. It doesn’t guess your acreage, invent a statistic, or fabricate a citation. Every program carries its official source link and the date we last checked it. That restraint is a feature, not a limitation — trust is the entire product.

Building it with an AI partner

Here’s the part that would have been impossible a couple of years ago. I’m not a software engineer. I’m the person who decides what the software should be and why anyone should care. Historically, that meant writing a spec and handing it off, then waiting, then negotiating over what got cut.

This time I built it directly, working with an AI coding partner. I’d describe the behavior I wanted in plain terms — “don’t call everything a grant,” “encode reality as a status instead of a fake date,” “flag when a family can’t apply directly” — and we’d turn each one into working software, review it, and tighten it. I made the product calls; the AI handled the implementation.

What surprised me is what didn’t change. The AI collapsed the cost of writing code to almost nothing, but it made the human judgment more valuable, not less. Every interesting decision was still a decision: how to score a match, how honestly to describe a closed program, how to word the eligibility flag so it’s clear without being discouraging, what to leave out of a first version. Taste, positioning, and knowing what the farmers actually needed — none of that got automated away. If anything, having a tireless builder made it more obvious that the scarce ingredient was a point of view.

I also insisted on the unglamorous discipline. We wrote an automated test suite so the core promises — a closed grant never shows up as the top result, a barn project returns financing by purpose and not a fake “barn grant” — are checked every time anything changes. I built an admin screen so program data can be corrected and verified through a form instead of buried in code, because the data has to stay honest as programs open and close. That’s the product-marketing instinct showing through: shipping isn’t the finish line, and credibility has to be maintained, not declared.

What actually shipped

Rural Funding Finder is live. You can type a project in plain language and get back real Michigan and federal programs, each one labeled by funding type, scored for fit, and attached to its official source and next step. There’s a detailed funding report for people who want the full application roadmap, a step-by-step assistant that helps you prepare an application without ever putting words in your mouth, and simple accounts so your profile and projects follow you.

I’m deliberately honest about the scope. It starts with Michigan and federal programs, not the whole country. The initial program data is structured and real, but it’s being verified against each official source before I’ll call it gospel — and the tool tells you which entries are confirmed. It’s a research head start, not a replacement for the agencies who actually run these programs. Saying that plainly is, again, the point.

What I’d tell another team

Three things came out of this that apply well beyond farming.

Positioning is a product decision, not a paragraph you write at the end. The single choice to ask “what are you trying to do?” instead of “which program?” shaped every screen that followed. Get that right and the interface designs itself; get it wrong and no amount of polish saves it.

Honesty is a feature you can ship. “Not everything is a grant” and “we won’t invent a deadline” aren’t disclaimers — they’re the reasons someone trusts the tool over the twelve open tabs. In a category full of vague optimism, being the one that tells you the truth is a durable advantage.

And AI changes who can build, not whether you need to know what’s worth building. The tooling let me go from insight to working product without a handoff. But it was the farmers’ frustration, and the discipline to keep the thing honest, that made it worth using. The technology got cheap. Judgment didn’t.

Try it: funding.altocitylimits.com

Rural Funding Finder is a self-service research tool from Alto City Limits. It provides the technology; each funding organization provides authoritative guidance. It doesn’t guarantee funding or eligibility — always confirm details with the official source before applying.