About
Someone has to make the rules usable
Regolyn exists because AI regulation arrived faster than most companies' capacity to read it, and the gap was being filled with fear, invoices, or nothing.
- Company
- Regolyn Inc., Delaware C corporation, United States
- Founded
- June 2024
- Headquarters
- Denver, CO, United States
- Stage
- Pre-Seed, $800,000 raised
- Team
- 12 people, US based
The story
Built by the people who used to be the workaround
Audrey spent 2024 answering enterprise AI questionnaires by hand at a logistics platform: three in one quarter, each a week of spreadsheet archaeology about tools nobody had inventoried. Daniel was building evaluation infrastructure that made language models reliable enough to trust with exactly that kind of tedious, high-stakes reading.
They incorporated Regolyn in 2024 on a simple bet: the coming wave of AI rules would land on companies that use AI, those companies would never hire specialists fast enough, and modern language models, properly caged, could do the mapping and drafting that makes readiness achievable without one.
The company is Pre-Seed, $800,000 from angels and operator syndicates, and the team of 12 works on this full time from Denver and remotely across the US. The founding team works on Regolyn full time.
Founders
Two people, one annoyance
Audrey Sloane
Co-founder and CEO
Audrey ran operations and vendor risk at a mid-market logistics platform, where she personally answered three AI questionnaires from enterprise customers in one quarter with no tooling. Regolyn is the product she kept wishing existed. Before that she led compliance operations rollouts across four countries.
Daniel Reyes
Co-founder and CTO
Daniel built ML retrieval and evaluation infrastructure at a data platform company, and spent two years making language model output reliable enough to bet a product on. He cares about one thing above all: systems that say what they know and ask when they do not.
Founder profiles are illustrative for this preview deployment.
How we work
Three habits we hire for
Plain language is a feature
If an operations lead cannot act on a sentence, the sentence is not finished. We edit until it is.
Grounded beats impressive
The engine cites its library, shows its reasons, and asks when it does not know. A confident guess is a bug.
Calm sells better than fear
Compliance marketing loves sirens. We publish effective dates and first steps instead, and let usefulness do the selling.
See where you stand this week
The free plan runs a real readiness check on two use cases: mapped obligations, a score, plain-language gaps, and drafted artifacts. Ten minutes of answers, no card.