Blog
Notes on evidence, diligence, and building Aventro.
Engineering·July 20, 2026·11 min read
The Hard Part of AI Due Diligence Isn't Intelligence. It's Trust.
Why we built Aventro around evidence, verification and traceability, not just better AI answers.
Read post
Perspective·June 23, 2026·1 min read
What a Daruma doll taught me about due diligence
A ritual I picked up in Tokyo is built entirely around focus. Venture diligence usually does the opposite.
Read post
Due diligence·May 6, 2026·1 min read
Most due diligence is just structured procrastination
We call it rigorous analysis. A lot of it is clerical work, and the deal does not wait for it.
Read post
Due diligence·May 4, 2026·1 min read
Your firm's memory is an asset. Most firms lose it.
Decades of hard-won judgment, sitting in PDFs nobody can search, or gone the day a partner retires.
Read post
Due diligence·May 3, 2026·1 min read
The quarter-trillion-dollar guessing game
Every year a fortune gets deployed into private companies on what still feels like guesswork.
Read post
Decision-making·May 2, 2026·1 min read
Conviction is not a gut feeling
The best meetings end in conviction. The honest question is how much of it is confirmation bias.
Read post
For founders·April 29, 2026·1 min read
Where early-stage founders actually sit on the map
Every founder claims to lead a blue ocean. To an analyst reading fifty decks a week, that is noise.
Read post
Risk·April 28, 2026·1 min read
What dead startups teach you
Ninety percent of startups fail. The useful lessons live in the ones that did, not the unicorns.
Read post
For founders·April 27, 2026·1 min read
Founders, stop letting diligence happen to you
You built something real. Then diligence stalls it for weeks while someone checks boxes and misses the point.
Read post
Risk·April 26, 2026·1 min read
Checked boxes don't stop failed investments
Deals rarely fail because a box went unchecked. They fail in the gaps between the boxes.
Read post
Perspective·April 17, 2026·1 min read
AI shouldn't make the decision. It should make it clearer.
Our whole approach rests on one line. Keep the human in the loop, and give them a sharper view.
Read post