Today we're launching Felix — the first platform that lets legal, finance, and insurance professionals automate their most complex operations without sacrificing reliability, auditability, or control.

We raised $1.7M in pre-seed funding led by XYZ Venture Capital, with participation from founders and executives at Amazon, Apple, Palantir, Flexport, Yelp, and Midjourney.

XYZ Venture Capital

Pre-seed

$1.7M

Founders and leaders from

Software is eating paperwork

Every advance in AI has made individual professionals more productive, but not businesses. The harder problem — turning the operational knowledge inside professional services firms into systems that scale — has remained unsolved.

Our belief: the next wave of automation will come from traditional software, not agents.

Traditional software is reliable, deterministic, and resilient — drawing on decades of security research and engineering investment. What has been missing is a tool for experts in professional services to build process automation software themselves, without an engineering team or lines of code written by hand.

We call this hyperautomation. It's what Felix is built to deliver.

How it works

Using Felix, professionals describe their process in plain language. Felix compiles it into real code that runs continuously without anyone prompting it — deterministic, auditable, and consistent at any scale.

Felix gives every professional services firm the software capabilities of an engineering team.

Here are a few early successes:

• Advocate, a New York risk management firm, had a ten-year backlog of unprocessed mortgage insurance compliance documents — 65,000 policies that needed to be matched, extracted, and verified. Felix turned that process into software that cleared the backlog in three weeks. It now runs continuously, and cross-referencing at scale surfaced discrepancies in supplier data that had never been visible before.
• Matthew Letts at Recourse Collective cross-referenced tens of thousands of creditor claims to surface phoenixing fraud — companies that dissolve and reincorporate to escape debt. Within eight weeks the system identified £1.4 billion in recoverable assets, with a full audit trail ready for regulators. One person, with Felix as his engineering team, completed work that simply wasn't possible before.

Why hyperautomation wins

For decades, the operations that define professional services — compliance reviews, claims processing, fraud detection, lending — have been executed manually, by skilled professionals, one case at a time. The software to do it reliably at scale has never been within reach.

Large language models promised to change that: AI that can reason through the reams of contracts, claims files, and regulatory filings that define this work. But for operational use at scale, LLMs presented hard limitations — run the same process twice and you get two different answers; outputs carry no audit trail; and routing tens of thousands of documents through a foundation model quickly becomes cost prohibitive. Workflow tools like n8n or Zapier are consistent and reliable, but built for simple triggers — not for compliance logic that pulls from multiple systems, applies years of regulatory expertise, and has to be right every time.

That gap is what Felix was built for. The product is live at felix.so.