Founder · BoardPath · Auris Intelligence
For years, Eric Tetzlaff jokingly told people he ran an adult daycare. It got a laugh every time — and it wasn't entirely wrong. He spent over a decade managing homeowner and condominium associations, where governance gaps have a way of bringing out the worst in otherwise reasonable people. When documents are ambiguous, precedent is tribal, and no one in the room can agree on what the rules actually say, behavior fills the void that structure should have occupied.
What Eric eventually understood — and what nobody in the industry was talking about — was that the behavior wasn't the problem. The information architecture was.
Over fourteen years managing hundreds of associations — as many as 34 at once — leading a 7-person team, and overseeing seven-figure operating budgets and $15M+ in reserve funds, Eric developed something unusual: a practitioner's intuition for how governing documents fail. Where authority hierarchies break down. Where amendment inheritance creates silent contradictions. Where the gap between what a corpus says and what a board needs to know becomes a liability — legal, financial, and operational.
That intuition became infrastructure when the right tools arrived. Eric was an early adopter of NotebookLM in beta — not as a productivity experiment, but as a deliberate attempt to solve a real problem he had been managing manually for years. He built siloed document corpora for each association he managed, constrained the system to cite only what existed, and — critically — gave it explicit permission to say the corpus does not address this. When it did, he didn't patch the gap with inference. He wrote amendments. He drafted board motion language. He built the documents that the system revealed were missing.
That feedback loop — retrieval surfaces the gap, human judgment closes it, corpus improves — became the architectural foundation of BoardPath, the governance intelligence platform Eric founded in 2026. BoardPath converts fragmented HOA and condominium governing documents into citation-grounded, meeting-ready answers, with a proprietary confidence scoring system that surfaces uncertainty, flags document conflicts, and never infers beyond what the corpus can defend. Where gaps exist, the platform doesn't simply report them — it characterizes them, flags their potential legal and operational risk, and offers boards a clear path toward resolution. The system that reveals what the corpus is missing becomes the system that helps close it.
Concurrently, Eric designed and built Auris Intelligence, a forensic legal document intelligence platform currently in active production use by a midsize regional corporate law firm for civil litigation defense preparation. Auris operates on a 110,000+ document corpus with SHA-256 chain-of-custody ingestion, a context-window-aware subagent scaffold designed to deliver signal-only inputs to an AI coordinator, and an output layer that produces attorney-ready case analysis, deposition preparation materials, and chain-of-custody attestations.
Eric's design philosophy is grounded in a conviction he arrived at the hard way: that the most important thing an AI system can do is know what it doesn't know — and say so clearly, before it says anything else. Transparency, retrievability, and honest uncertainty aren't features to be added after the architecture is designed. They are the architecture.
He builds systems that work the way good judgment works: from the ground up, stress-tested until they break, rebuilt until they don't, and handed to the people who need them with a clear account of exactly what they can and cannot be trusted to do.