Why I built Mavenly — and the problem I could not stop thinking about.

I did not set out to build grant software. I set out to solve a problem that kept showing up in every organization I worked with — a problem nobody described the same way twice, but everybody recognized instantly once you named it.

Here is the shape of it. A nonprofit doing genuinely excellent work applies for a grant it should win. It loses. Not because the program is weak, not because the need is unproven, and not because the funder was hostile. It loses because the application was assembled in the last eleven days by three people who were already at capacity, pulling numbers from four spreadsheets, reusing a narrative written for a different funder with a different theory of change, and answering a budget question at 1 a.m. that deserved a week of thought.

The work was never the problem. The scaffolding around the work was.

I spent years in systematic trading before any of this — building software that had to make decisions under deadline pressure, with real consequences, using incomplete information. That world taught me something I have never been able to unlearn: when smart people consistently produce bad outcomes, you are almost never looking at a talent problem. You are looking at a process that forces them to make their most important decisions at the worst possible moment, with the least available context.

Grant funding is exactly that process. And it is applied, disproportionately, to organizations with the fewest resources to absorb it.

Through T. James Enterprises and the philanthropic work under it — HBCU Endowment Fund, World Nourishment Foundation — I saw the same pattern from the funding side. Development directors who were extraordinary at their jobs, spending sixty percent of their week on document logistics. Institutional knowledge that lived in one person's inbox. Renewal deadlines discovered by accident. Reporting obligations that arrived like weather.

So the question I started with was not 'how do we help nonprofits find more grants?' That is the question most tools in this category answer, and I think it is the wrong one. Discovery is rarely the binding constraint. Most organizations already know more funders than they have the capacity to pursue properly. Handing them a longer list does not help; it just moves the bottleneck one step downstream and adds guilt to it.

The real constraint is throughput with quality. How many well-scoped, well-evidenced, genuinely aligned applications can a three-person development team put into the world in a year — without burning out, without recycling narratives that no longer describe the work, and without letting a renewal quietly lapse?

That is what Mavenly is built to raise.

Which is why the product is shaped the way it is. Matching comes before search, because relevance beats volume. The proposal editor is grounded in the organization's own library — its evaluations, its budgets, its prior awards — because a draft that invents evidence is worse than no draft. The pipeline, the deadlines, the reporting calendar, and the compliance rollup all live in one place because in practice they are one workflow that has been artificially split across four tools and one person's memory.

And AI sits underneath all of it as a layer, not a gimmick. Maven drafts, summarizes, flags risk, and explains its reasoning with citations back to your own documents. It does not write the sections that require lived experience or community context — we deliberately hold it back there, because those are the paragraphs where a plausible-sounding fabrication does the most damage to trust that took years to build.

There is a version of this company that optimizes for the funders. That is not the one I wanted to build. Foundations are not adversaries — they are infrastructure, and good ones are doing careful work under their own constraints. But the asymmetry in the relationship is real: the applicant carries almost all of the process cost. Mavenly exists to move that cost off the applicant's desk without lowering the quality of what the funder receives. Done right, both sides win, and the money reaches the work faster.

The organizations I have in mind while building this are specific. HBCUs with endowment offices of two. Food security nonprofits operating across four counties. Community health groups whose outcomes data is real and rigorous and trapped in a PDF from 2023. They do not need to be told to be more strategic. They need the fifteen hours a week that the paperwork takes.

That is the whole thesis. Give capable people back their time, keep the quality of the work intact, and let more good proposals reach the funders who were always going to say yes to them.

We are building it in public, shipping weekly, and writing here about what we learn. If you are doing this work and something in the product gets in your way, tell us. That feedback is the fastest path to a version of this that actually earns its place in your week.