Building a Tech Company in Maine

George Matelich is the co-founder and CEO of Rely, a Portland-based AI platform that automates due diligence for multifamily real estate transactions. The platform reads leases, rent rolls, contracts and financial records across an entire data room and links every finding back to its source document, a feature Matelich describes as visual grounding. A process that typically takes a team of three to seven people five to ten business days can, with Rely, be completed by one person in an afternoon. Co-hosts Dr. Lisa Belisle and Dr. Jeff Barkin drew Matelich out on why that traceability is the foundation of the product: in a field where AI skepticism runs high, giving users a direct path back to the underlying document is how the software earns the right to do more.

Much of the conversation turned on what it takes to build a company in Maine rather than in a coastal technology hub. Matelich was candid that access to venture capital and to dense professional networks is harder from Portland than from San Francisco or New York, but he pushed back on the premise that geography is destiny: his customers are concentrated in the Southeast, Southwest and Midwest, not in the major tech corridors, so proximity to those markets matters more than proximity to any particular investor community. He is also the co-founder of Nexus Maine, a statewide community connecting entrepreneurs, and he and the co-hosts discussed how informal networks of founders, funders and mentors can surface talent and opportunity that would otherwise stay invisible across a dispersed state. Matelich was equally direct about the psychological weight of founding: a long discovery period, a high rate of early failure and a tendency to tie personal identity too closely to professional outcome are real hazards he said he has not fully solved.

Barkin framed a question the full episode returns to repeatedly: what does widespread AI adoption actually look like for people who have never tried it, and what happens when trust in these tools goes wrong in either direction. Matelich argued that the same principle guiding Rely, confining the model to a finite, verified body of documents and building in a mechanism for users to push back on any output, applies broadly to responsible AI deployment. Belisle extended the parallel to healthcare, where the stakes involve lives rather than transactions, and where clinician and patient skepticism of technology-driven tools remains substantial. The three agreed that earning trust through low-stakes trials, maintaining human oversight and resisting the impulse to take too large a leap too soon are the conditions under which any high-stakes AI application can credibly advance.