Solo Founder Project · Early Client Work
AI tooling for nonprofit grant operations, built out of a real frustration I watched up close.
Growing up, a nonprofit was a material part of how my family navigated a really difficult stretch. My childhood was shaped by PANDAS/PANS, a neurological condition that took years to diagnose and manage. The organizations that showed up during that time were run by people who genuinely cared but operated on tiny budgets, constantly chasing funding just to keep their programs alive.
I stayed connected to that world. I co-founded the National PANDAS/PANS Youth Alliance, served as a board ambassador for the Northwest PANDAS/PANS Network, and did state-level nonprofit advocacy in Washington, California, and Vermont. Through that work, and through SOURCE Nonprofit Consulting, where I led a workshop on AI in grant writing for 30+ nonprofit leaders, I kept running into the same thing: executive directors spending 30–50% of their time on grant operations instead of their actual mission.
“Our target accounts are wasting enormous time finding, tracking, applying to, and reporting on grants, instead of working on what they’re passionate about in the first place.”
Grant operations is a second full-time job nobody signed up for. Small and mid-sized nonprofits sit in a bind: they require grant funding to survive, but they don’t have the time, people, or resources to stay on top of it.
What I’ve heard repeatedly, in the SOURCE workshop and in conversations since, is that nonprofit leaders have to wear multiple hats at once and can’t put their full effort into grant research, applications, and management.
Tailored agentic workflows built per client. Right now Grantara is AI workflow, tooling, and visualization products, built on a set of agentic workflows and dashboards that handle the most time-consuming parts of grant operations. My north star for this work is simple: provide value for my clients, whether that’s a grant that wasn’t on their radar or helping them manage every application in one place.
Continuous scraping of Grants.gov, Candid, foundation websites, and state portals, with searches optimized per client. Results go into a daily, weekly, or as-needed digest of relevant opportunities scored against the org’s mission, budget, and geography, so the EDs I work with aren’t manually sifting databases.
For each opportunity, a scoring layer pulls funder history — who they’ve funded, at what amounts, for what program types — and ranks fit probability. The goal is helping orgs skip the applications that aren’t worth their time. This is one of the hardest problems I’ve worked on, since the scoring needs client-specific criteria to be genuinely relevant.
For greenlit opportunities, the system drafts full application narratives using the org’s program documentation, theory of change, and past grant language as context.
For each client, deadline tracking, submission status, reporting due dates, and funder communication are organized together the way the ED wants, so they have one hat to put on instead of five.
How I validated the problem before building. The SOURCE Summit workshop was my biggest proof point. I led a session on AI in grant writing for 30+ nonprofit leaders and spent as much time listening as presenting. From those conversations, people weren’t struggling with the concept of AI helping with grants. They were unsure how to implement it in practice while maintaining a human level of quality in research and writing.
Services-first, not SaaS-first. The SaaS layer is already consolidating, so building another general product didn’t make sense. By doing specific grant-related work for each client using AI workflows and products, I actually do the work, rather than selling access to something that still requires them to do the work in a slightly different way.