Our Thesis

Analysts, associates, and principals will spend less time manually reading data rooms, rebuilding cap tables, copying figures, and drafting first-pass memos. Their edge will come from something sharper: asking better questions, reviewing AI-generated work with discipline, challenging assumptions, and presenting a clearer investment case.

Venture diligence is still painfully manual. Research on VC decision-making suggests investors spend an average of 118 hours evaluating a startup before investing, with some firms spending several hundred hours. A memo process that used to take 80–120 analyst hours over two to three weeks can now be compressed into roughly 20 minutes.

That is not a 2x or 10x improvement. It is closer to a 240x–360x compression in manual memo preparation time.

But speed is only useful if it takes you in the right direction. This is the part most AI tools still miss. They can summarize a clean deck… but they struggle when the real diligence work sits across hundreds of files — messy PDFs, cap tables, and financial statements, where the challenge is not just reading the data once, but recalling the right detail accurately.

This is why we built Mentis.

AI tools have been hallucinating for far too long. Investors cannot rely on that when real capital is on the line. Mentis is built to bring that trust back: a diligence partner built for accuracy, recall, and speed.

Less time buried in manual work, more time focused on the parts of diligence that actually matter.

Speed usually comes at the cost of depth. With Mentis, it doesn’t have to.