Tell us about your company and the problem it solves, or its benefit to learners or educators.
BoodleBox is a collaborative AI workspace built for teaching, learning and work. Nearly every AI tool on the market is single-player: one person, one prompt, one private conversation that disappears when the tab closes. We built BoodleBox around a different idea: that thinking gets better when people do it together and that AI should amplify human judgment rather than stand in for it. For students, that means learning to work critically with multiple AI models rather than simply asking one chatbot for an answer. For educators, it means bringing AI into a course with real guardrails and visibility into how students actually use it. For institutions, it means secure access that keeps them in control of how AI shows up on their campus and a workspace where the thinking people do together builds into something the institution keeps. That’s the infrastructure institutions need to become both human-first and AI-native.
Where did the idea come from to create your company?
BoodleBox comes from three convictions I’ve carried my whole life: service, education, and collaboration. They’re also this company’s core values and that wasn’t an accident.
I came to this country as a refugee. First a plane, then a tent, and somehow the American Dream. Every door that opened for me was opened by education. Every door that opened for me, education opened: West Point, the White House, a combat tour in Afghanistan, a university trusteeship, and a few companies along the way.
West Point is also where the AI problem came into focus. Colonel Chris Lowrance, a professor there, framed something I haven’t been able to unsee. For most of history, the hard part of any task was the middle — generating the solution — and schools and companies built themselves around training people for it. AI does that part almost for free now. So the work inverted. What’s left is at the two ends: framing the right problem going in, and validating the answer coming out. A valley where a hill used to be. It’s called the Lowrance Curve, and those two ends are discernment, which happens to be the things our institutions practice least.
Meanwhile I watched the most powerful technology in human history get deployed in a way that isolated people, rewarded the already-privileged, and confused producing an answer with solving a problem. Students and faculty each disappeared into private conversations with a chatbot, and institutions had almost no visibility into any of it.
That crystallized BoodleBox. AI isn’t a faster horse — it’s a car, something genuinely new that can reach places no horse could. But a car is only as good as the road beneath it. Everyone is handing out faster cars; almost no one is building better roads. So rather than build another chatbot, we set out to build the road: the shared space where people frame the problem and validate the answer together, generation handled by AI, judgment kept human. I laid the full argument out publicly in a manifesto, The Better Roads Manifesto, which readers can view here. It includes visuals of the Lowrance Curve.
Tell us about one challenge and how you overcame that challenge.
Our first real obstacle wasn’t skepticism about AI. It was fear of it. We’d walk into a faculty meeting and the first question was never “how does this work” — it was “how do I stop students from cheating with this.”
The market’s answer at the time was detection: run the submission through a checker and get a verdict. We thought that was the wrong instinct, and worse, an unwinnable one. Detectors can’t reliably tell honest work from misconduct, so faculty were being handed a tool that produced accusations they couldn’t defend. Some responded the only way left to them and went back to blue books. One university system handed AI to half a million students with no training and got exactly that — panicked faculty and frustrated students.
Our position was that the problem wasn’t the finished document, it was that faculty could only ever see the finished document. They were being asked to judge a process they had no window into. So we built for visibility instead of detection: an instructor can design an assignment with guardrails, see how a student actually worked with AI, and grade the thinking rather than guess at it.
That was a harder sell in the short term, because “we’ll catch them” is a much simpler promise than “we’ll show you the work.” It took sitting with faculty who had already been burned. But it’s the only version that survives contact with a real classroom, and it’s why the conversation with a department chair now starts somewhere other than fear.
What are you most proud of or what is your company’s greatest achievement?
In a short time we’ve shown there’s real demand for a human-first approach to AI. We raised $5 million after hearing no from 137 investors, and reached $2 million in annual recurring revenue in under 18 months. Those numbers matter to me because of what they represent: educators, learners, and institutions choosing a different model for AI, one at a time, against the grain of what the market was telling them to want.
What I’m proudest of is narrower than that. It’s watching a student learn to frame a problem before reaching for a model, interrogate the answer that comes back, and recognize the point where their own judgment has to take over. One professor told us she’d read the best thesis statements she’d seen in two years. That’s the whole thing. Not students producing faster, students thinking better.
If we help people learn to use AI critically and responsibly rather than just quickly, that’s the achievement that lasts.
Where do you see your company in five years?
Five years out, the question institutions are dodging today will be unavoidable. Careers used to start on the generation rung — you produced for years, and the doing of it taught you to judge. AI has taken that rung. Every institution will have to decide how it develops people when the entry-level work is automated, and the ones that started building for discernment now will be several years ahead of the ones that waited.
I don’t think the answer looks like one model inside one app. People will work across an ecosystem of models, and institutions will need environments that help them choose well while keeping transparency, security, and accountability intact. That’s the infrastructure we’re building — for education first, and increasingly for workforce development, because the fluency a student builds in a sophomore seminar should still be theirs, and provable, at their first job.
The technology will change fast. What I want is simpler than a market position: working with AI should feel as ordinary as brainstorming with a colleague does today, and just as clearly something you did, not something that was done for you.