Unscripted SaaS · Founder stories
Cracking Distribution Is The Hard Part
with Tom Girgash — founder of Taproot, the memory layer for solo AI operators
Tom Girgash is in his mid-twenties, graduated from Ohio State, took a sales role selling bearings and hoses, realised he was more interested in systems and software, and is now doing a master’s in information systems at Cincinnati while building Taproot. When host Jeremy Rivera asked what problem he has not solved yet, the answer was not technical.
Marketing At Scale Is The Toughest Challenge
“I mean, I know we just talked about it, but I think, you know, it’s only been a couple of weeks, but I think cracking distribution and marketing has been, is going to be the toughest challenge and so far it is, you know, the toughest challenge to do at scale. You know, I’m not in a position right now I want to spend money on ads or, you know, anything of that nature. So, I mean, trying to figure out a good way to create an organic distribution engine is something that’s interesting to me. And it has been, you know, rather difficult.”
— Tom Girgash, TaprootThe honesty about the timeline — “it’s only been a couple of weeks” — is worth keeping, because it stops this being a war story from someone with the answer. He is describing the problem while inside it, which is rarer and more useful than the retrospective version.
His constraint is the one most bootstrapped founders actually face: paid acquisition is off the table, so the only available channel is one that has to be built rather than bought. And an organic distribution engine has an unhelpful property — it does not respond to effort on the timescale that a runway does.
His current traction is about twenty signups, which he calls “pretty good for someone who’s only been doing it for a couple weeks,” noting it is Mac-only for now because he is a native Mac user and it was the fastest way to ship.
Building Was Not The Bottleneck
The striking thing about the interview is how quickly he moves past the technical work. The product connects a personal knowledge base to any AI tool that speaks MCP — “Codex, ChatGPT, Claude, Claude Code can all connect to it” — which he considers the point: “you have a central knowledge base of your information, your knowledge, the things you want AI to be able to tap into.”
His thirty-second pitch is clear enough:
“Taproot gives the ability for people to connect their knowledge with AI. Everyone has different contexts, but if you’re a heavy AI user, sometimes the reasoning you get back from AI is not actually rooted in your work. So as you’re working with different clients in different contexts, it can get a bit confusing about where different clients stand or different projects stand.”
— Tom Girgash, TaprootBuilding that was tractable. What is not tractable is the messaging, and he is disarmingly direct about it: “Still trying to work on the messaging myself. It’s been one of the harder parts, I will say.”
His test for it is the one that should be in more founder toolkits. He tried the family version — “like the mom test… your mom will always say it’s a good idea. But my mom didn’t understand what it was, and same with my sister.” The failure was not enthusiasm. It was comprehension, which is a much more actionable result.
Building For People You Actually Are
Asked why anyone should trust him as a builder, Tom’s answer was about proximity to the problem rather than credentials:
“The reason I’m building it is because I’m trying to solve my own problem. I’m not trying to build something just for a random group of people that I really don’t know much about. It’s for solo operators, people like myself that are using AI a lot, dealing with different contexts — running the issues of AI not remembering the things they’re working on, as well as not having a place where your knowledge lives.”
— Tom Girgash, TaprootHe is honest that the sample size is one: “I mean, it’s only myself, but I think solo operators, in the case of AI, it’s probably going to become a bigger thing than it is now.” That is a bet on a market forming, not evidence that one exists — and he presents it as such.
There is a real advantage in it. Building for yourself gives you a feedback loop with no latency and no research budget. There is also a real risk, which is the messaging problem he already hit: when you are the user, the vocabulary that describes the product is your internal vocabulary, and it does not survive contact with anyone outside your own workflow. His mother and sister found the boundary of it immediately.
He has an idea of where it goes next — “a centralized knowledge layer for a firm or a team, a small team,” so people could see what others are working on — but is deliberately parking it: “that’s definitely something more down the road. For now I’m more focused on solo operators.”
Finding The First Hundred Solo Operators
Practitioner guidance from us. Tom describes the difficulty; he does not offer a playbook, and we are not attributing one to him.
The specific hardness of his problem is that solo operators have no shared address. There is no conference, no job title, no procurement department. They are distributed across communities organised around other things entirely.
Jeremy offered the most concrete lead in the conversation, describing where the practitioners went after the platform churn of recent years: roughly a quarter to Threads, half to Bluesky, and a quarter into Slack communities. His note on the last group is the useful one — they are “harder to get into, but great for mining information if you can find active Slack communities of people that have a similar interest.” He also flagged Skool as worth investigating for free and paid communities.
The pattern worth taking: for a product whose users have no category, the distribution work is not broadcasting into a channel. It is being present, usefully and repeatedly, in the small rooms where those people already solve the problem manually.
What The Constraint Really Is
Tom’s own read on the macro picture lands in the same place as several other guests on this show, and he arrived at it from the builder’s side rather than the marketer’s:
“As software at least becomes a bit commoditized, distribution and marketing is going to be — that’s going to be the most important aspect to building a business, software business, name any business really. Being able to market your application, but I think also because AI slop, people get tired of that — being human, being as human as possible, connecting with your audience on a very human level, it’s going to be something that’s going to be pretty large going forward.”
— Tom Girgash, TaprootWhich puts him alongside Nick Eubanks on distribution as the remaining moat — a point Jeremy raised in the conversation itself. The difference is that Tom is not describing a strategic principle. He is describing the thing currently standing between him and a business.
Taproot is at taproothq.com. Tom is on LinkedIn as Thomas Girgash and on X as @TGIRGS.
More builders at this stage in founder stories and bootstrapping, and the onboarding side in product-led growth.