The analyst seat on your team. A company built by one ex-analyst.

I trained as a software engineer before switching into finance, where I stayed for a decade. I ran corporate-finance and capex analysis for Fortune 500 management, worked as lead analyst on M&A deals that closed, and finished as a principal at a family office deploying over $100M into private equity and venture across banking, real estate, and technology. The roles kept changing. The core never did: build the models leaders rely on to decide, and make sure every assumption is grounded and the memo holds together.

Here's what that decade really felt like. The work is hard in ways the tools never acknowledged, and for most of those years the tools were just Excel and PowerPoint. The data is messy and arrives in a dozen shapes that refuse to line up. Research eats days. Modeling is slow, because a model is just the logic you write out by hand in Excel, always against a deadline. And the people you do it for are demanding because they have to be. A senior leader is about to make a call worth more than you'll earn in a decade, and they're making it on your numbers. The standard is unforgiving, as it should be, because a wrong figure becomes a bad decision.

Now look at it from the other chair, the half nobody talks about. The leader making the call is just as stuck. They can only decide as fast and as well as their analysts let them. Good analysts are expensive, and it takes years to train one to truly understand an industry. There are only so many hours in a day. So a hundred-million-dollar decision ends up capped by the bandwidth of a few tired people working against the clock. That's the real problem, and it sits with neither the analyst nor the leader. Rigorous financial work simply doesn't scale.

The tools never helped with any of this. Excel is where the work happens, and it's bad at most of what you ask of it. The shared drive where all that work lives is worse. And the new wave of AI-for-finance tools manages to be worse still. They'll write you a forty-page memo in an afternoon, and the structure looks right and the prose sounds confident. But the assumptions come from nowhere, the numbers don't tie, the citations are invented, and if you run it again you get a different answer. It looks like an analyst's memo without being anything like one. Enough of them are shipping now that the work itself is losing credibility inside the firms that used to take it seriously.

I built Kennan to do the hard part instead of dodging it. It isn't another tool laid on top of the old ones; it's meant to be the analyst itself — an agent that produces the work a real analyst ships, held to the standard the work has always demanded.

It runs on a few convictions that are unfashionable right now. Every number it gives you traces back to where it came from. Run the same analysis twice and the answer doesn't move. Every version can be locked, with a record of who changed what and when, so the work still holds up when legal reviews it or someone reopens the deal a year later. And your documents stay yours, handled under signed agreements with no data retention and EU residency if your firm needs it.

I'm building this for institutions rather than consumers: corporate finance teams, asset managers, private equity firms, anyone who has to defend a business case in front of a board or an investment committee. It's not a chat assistant playing financial advisor.

If your firm cares more about whether the work holds up than how fast the reply lands, and about analysts keeping authority over the work that carries their name, I'd like to talk.