Reg42 for investors

Financial firms spend close to $1 trillion a year on compliance.1 Reg42 turns that work into code.

~$1Ta year on compliance at financial firms.1
417,070US compliance officers, the people we build for.3
42%of C-suite time at large US banks went to compliance in 2023, up from 24%.4
45%of practitioners rate their approach effective, down from 59%.5

What problem do we solve?

Three pains every financial compliance team has:

  1. No certainty that every applicable obligation is addressed. A material inherent risk.
  2. Heavy dependence on manual work, by internal and external teams. An efficiency killer.
  3. Long time to market for new products, services and markets. A growth blocker.

What does Reg42 do?

We automate a financial firm’s compliance program with vertical AI and harmonize it in one system, to reduce risk, raise efficiency and speed up growth. It runs on CLHEAR, the first regulation-as-code standard for the whole compliance program, and Solon, our AI, which answers only with evidence.

Why now?

AI now does expert work. The tasks the best AI agent can finish grew from about 11 minutes of an expert’s time in June 2024 to more than 16 hours in April 2026.6 Law already has its own AI: in a 2025 test, legal AI products beat lawyers on research accuracy.7 Financial compliance has no standard and method for this yet. Someone will set them. We intend to.

How big is the opportunity?

Close to $1 trillion a year,1 including $206.1 billion on financial crime compliance alone.2 We start with the chief compliance officer’s program.

What is the moat?

CLHEAR, our patent-pending language (US application 64/161,186), the rules encoded in it, and each client’s compliance graph. Over time, anonymized data in CLHEAR’s format will show what is good enough for each company profile: a standard no one else will have, if we grow fast enough.

What is live today?

A working prototype, not yet a production product. On a fictional US broker-dealer with synthetic records, it shows what applies, the policies, what the firm runs, the gaps between them, and Solon’s answers with their evidence. We show it in a demo on request.

How does Reg42 make money?

A paid pilot first. It then becomes an annual subscription, priced by the scope of the client’s compliance program.

Who else is doing this?

To our knowledge, no company closes the full loop yet: map, compare, comply. Others cover a part: AI search, point tools, rule feeds, GRC registers. Reg42 joins the whole program, with its evidence. Where we don’t yet offer an alternative, we connect to a firm’s existing internal and vendor systems.

Who are the founders?

Avner Yoffe, CEO. 10+ years in compliance technology. Led the creation of eToro’s regtech department; earlier at Cappitech, now part of S&P Global. Author of CLHEAR.

Eitan Ezra, CTO. 6 years building mission-critical intelligence systems, leading software teams.

Want to dive deeper?

The thinking behind Reg42, with dated predictions you can hold us to.

A letter from our founder

“The intelligence has arrived. Nobody has onboarded it.”

Avner Yoffe, Co-founder and CEO

Read “Don’t Panic. Regulation is Cool.”

1. Reg42 estimate. TheCityUK and PwC put regulatory compliance at more than 13% of UK financial firms’ operating costs (November 2025). Applied to the operating costs of banks, insurers and asset managers (McKinsey, Swiss Re and BCG data), and adding other financial firms, that comes close to $1 trillion a year. 2. LexisNexis Risk Solutions, True Cost of Financial Crime Compliance, 2023, a survey of 1,181 professionals. 3. US Bureau of Labor Statistics, May 2025. 4. Bank Policy Institute survey of 20 member banks, October 2024. 5. CUBE, From adoption to assurance: the cost of compliance in an AI-enabled world, a survey of more than 2,000 practitioners in 11 countries, November 2026. 6. METR, Time horizons: the length of software tasks, in human expert time, that the best AI agent completes half the time. The April 2026 figure is beyond what METR’s tasks can reliably measure. 7. Vals AI, October 2025: 200 US legal research questions put to three legal AI products and to lawyers from a US law firm. On accuracy the products scored 78 to 81% and the lawyers 71%.