Dillon Auxier and Brandon Dey did not build Plato Industries around venture capital. Instead, the co-founders describe a customer-funded model where clients co-own the resulting intellectual property through a joint venture go-to-market structure, a very different path from the typical Silicon Valley AI startup playbook. Their focus is small and mid-sized businesses, along with lower-middle-market industrial private equity firms, groups that rarely get custom AI tooling built for them.
The conversation gets concrete fast. Auxier and Dey describe building an AI-driven workflow product for a local law firm handling roughly 2.2 billion in transaction volume, and another project built directly for Brad's own Medicare agency to track commissions automatically. Those two examples do more to explain what Plato actually does than any abstract pitch could.
Demystifying How AI Actually Works
Much of the episode functions as an accessible technical explainer, walking through how large language models work as next-token prediction engines, the differences between supervised, unsupervised and reinforcement learning, and what the industry calls the DeepSeek moment, when a lower-cost model challenged assumptions about what it costs to train frontier AI compared to OpenAI's approach. It is a rare, deep dive for a Sarasota business podcast, but Auxier and Dey keep it grounded in plain language throughout.
Building a Model for Trey's Roofing Business
One of the most tangible threads is a live example: Plato is building a predictive model aimed at Trey's roofing business to help identify hail-damage claims in the Dallas-Fort Worth market, turning an abstract AI conversation into a project the hosts have a direct stake in.
Roundabouts and Rapid Fire
The show's usual roundabout question and Rapid Fire Gauntlet close out the episode, though the conversation's real weight sits earlier, in how Auxier and Dey are quietly building practical AI tools for Sarasota-area businesses that would otherwise never have access to that kind of technology. Their customer-funded approach also offers a template other founders in the audience could realistically follow, without needing the venture capital connections that most AI startup stories assume are required, making Plato's Sarasota approach worth watching as the technology keeps evolving.
What you'll take away
- Plato Industries uses a customer-funded venture model where clients co-own the resulting IP rather than relying on traditional VC funding.
- The company built an AI workflow tool for a local law firm processing roughly 2.2 billion dollars in transaction volume.
- Plato built a commission-tracking tool for Brad's own Medicare agency.
- The founders explain large language models as next-token prediction systems and discuss the DeepSeek moment versus OpenAI's approach.
- Plato is building a predictive hail-damage model for Trey's roofing business in the Dallas-Fort Worth market.