The Finance Startup Bringing Agentic AI to Wall Street

In the last seven days, we signed as many contracts as in the whole of quarter four. There is clear, tangible value, and it is accelerating.

You have to be passionate and persevere. If that sounds like you, build a startup.

We are here with Arnie and Chas Englander, founders of Model ML from Winter twenty-four. Before this, they built and sold Fancy and Fat Lama. Welcome back.

Model ML is an AI workspace for financial services: our versions of Word, PowerPoint, and Excel on top of an agentic system connected to your files, email, CRM, data vendors, and public data. It cuts the gathering and analysis grind. Traction-wise, we are signing fast and are now used by roughly ten percent of the largest private equity firms and investment banks, plus asset managers and funds.

What were teams doing before, and why did you build this?

People lived in Office and Outlook, doing repetitive work. After we sold our last company, we tried angel investing, realized we were bad at it, and built a tool that auto-created a one-pager by pulling LinkedIn, Crunchbase, public filings, reviews, and more—what a human would do first. Others wanted it; we were better builders than investors, so we leaned in. Earnings summaries. Analysts used to spend days stitching filings with data like consensus from vendors, checking every figure. With Model ML, a connected spreadsheet and template auto-generate the pack into SharePoint or Google Drive—ninety to ninety-five percent done, often more accurate because we cross-check multiple sources.

How have the models and the market evolved in the last year?

Last year was testing; this year is using. Function calling and overall reliability improved, so even if we did nothing, our product would still get better. For data extraction from filings, models are already beating humans at top firms, so lower-level gathering and presentation are being automated.

Vision models plus OCR changed the game—reading tables and charts accurately unlocked a lot.

It is CEO-level or the most senior leaders. We build trust with real-data, laptop-open demos, lots of face time, and customer-specific work. We are on the ground in New York, London, Hong Kong, Singapore, and India.

What lessons from Fat Lama and Fancy shaped how you build now?

Perseverance—logical, not blind. Be calm through surprises. Work ethic matters; our team currently works six days. Obsess over customers; doing the first fifteen hundred deliveries ourselves at Fancy hardwired that.

It is a ridiculous rollercoaster. Hire for culture and joy of working together, not just CVs. We handled disasters—Stripe froze payments, so we took phone, cash, PayPal—do whatever it takes.

I still sit with users during trials, building and selling together. The next shift is full autonomy: tasks will run overnight without anyone clicking Run, so parts of the UI will matter less.

Ship faster to learn faster. We aim to stay extremely lean.

What keeps you motivated, and what would you tell a twenty-one-year-old who wants big impact?

We want impact and delight—those wow moments are priceless. Go all in if you love building. Worst case, you learn a tremendous amount; the jobs will still be there.

Startups are hard and long. If passion and perseverance fit you, start; if not, join a small team to get your work into users’ hands quickly, then jump later.

What should finance hires unlearn when joining a startup?

Unlearn big-company habits, think from first principles, and build the plane while rolling down the runway.

Why do YC again, and what was the Winter twenty-four AI batch like?

YC sets a cadence—weekly, daily, even hourly—and gives an instant network you cannot replicate, especially coming from Europe where seven-day work is uncommon.

San Francisco’s AI buzz was electric—working side by side with builders, it felt like constant breakthroughs. We love San Francisco. London can work, but it is harder to find extreme work ethic. Europe has great engineers; the Bay Area is expensive with fierce competition.

Roughly eighty percent of our customers are in the United States. I spend most time in New York; we also cover London, Hong Kong, and San Francisco. Many global tech decisions are actually made in San Francisco. If you can, move there; otherwise, be in a tier-one city close to customers.

Thanks for coming back to YC—great to see you both, and thanks for watching.