Founder's brief · 90 sec · Why compliance needs a twin, in Richard's voice
For compliance & regulatory change leaders at banks

The digital twin of your compliance function.

RegTwin keeps a working model of your obligations, controls, and policies. When a regulator publishes something new, eight agents work out what it touches, draft the changes, and put the package in front of your team for approval. Every step is logged as it happens.

AGENTS RECOMMEND · HUMANS APPROVE · EXAMINERS SEE EVERYTHING

Supervisory event traceREPLAY
09:41:07Monitoring · Agent 01New issuance detected: FRB SR letter, third-party risk management. Classified: applicable.
09:41:32Mapping · Agent 03Mapped to obligation register. 3 controls impacted: vendor onboarding, ongoing monitoring, exit planning.
09:42:15Drafting · Agent 05Policy amendment drafted with redlines. Gap memo prepared for second-line review.
09:42:18Evidence · Agent 08Evidence pack assembled: source text, mapping rationale, draft lineage. Audit-ready.
11:06:44Head of Reg Change · HumanReviewed. Amendment approved with one edit. Routed to policy owner.
ELAPSED: 85 MIN, ISSUANCE → APPROVED DRAFT✓ FULL LINEAGE RETAINED
Why this exists

Regulatory change has outgrown the process for handling it.

At most banks the process is still manual. Someone reads the issuance, decides what it touches, and updates a spreadsheet. That worked when there were a few changes a month. At today's volume it doesn't, and the gaps surface in exams.

200+

Regulatory alerts a compliance function can face on an average day, across jurisdictions and agencies.

THOMSON REUTERS · COST OF COMPLIANCE

60%

Of compliance officers who expect their personal liability to increase while headcount stays flat.

INDUSTRY SURVEYS, REG CHANGE FUNCTIONS

Weeks

Typical lag between an issuance landing and the control framework reflecting it. Examiners ask about this gap.

THE GAP REGTWIN CLOSES

The lifecycle of a reflex

How a regulatory change moves through RegTwin.

Five stages. Each one's output feeds the next, and the full trail is kept.

01

Sense

Agents monitor the agencies you answer to: SR letters, bulletins, final rules, enforcement actions. Each item gets classified for whether it applies to your charter and footprint.

Agents
02

Map

Applicable changes are traced through the twin to find which obligations they alter, which controls they touch, and which policies now have a gap.

Agents
03

Recommend

The system drafts redlined policy language, control updates, and a gap memo, with the reasoning and source citations attached.

Agents
04

Approve

Nothing takes effect until someone on your team reviews it, edits it if needed, and signs off. RegTwin does the reading and the drafting. The decision is still yours.

You
05

Evidence

Everything lands in an evidence pack: what was detected, how it was mapped, what was drafted, who approved it, and when.

System of record
The docket

Eight agents, each with one job.

Each agent handles one part of the process and hands off to the next. Narrow scope keeps the outputs easy to check, and easier to explain to your model risk team.

AGENT 01

Monitoring

Watches the relevant regulators and classifies new issuances for whether they apply to your institution.

AGENT 02

Interpretation

Reads the full text and extracts the operative obligations: what has to change, by whom, by when.

AGENT 03

Mapping

Maps each obligation to your control framework and policy inventory and flags every touchpoint.

AGENT 04

Impact assessment

Scores impact and urgency against your risk appetite so the queue arrives ranked.

AGENT 05

Drafting

Produces redlined policy amendments and control updates, with rationale and citations attached.

AGENT 06

Challenge

Reviews the other agents' work adversarially before it reaches you, the way a second line of defense would.

AGENT 07

Drift monitoring

Watches for drift after adoption, meaning controls that no longer match the obligations behind them.

AGENT 08

Evidence

Keeps the full lineage of every detection, draft, and approval, packaged for examination.

There is no setting that lets RegTwin change a policy on its own. Every output is a draft until someone on your team approves it.

The operating doctrine
We don't sell autopilot, because no examiner would accept it and no honest vendor would claim it. The agents do the reading, the mapping, and the drafting. Your people make the decisions, and the record shows who decided what.
§

Built to survive SR 11-7 scrutiny

We expect your model risk team to validate RegTwin like any other model, and the documentation for that review already exists.

§

Every recommendation shows its work

Source text, mapping logic, and draft lineage travel with each output, so there are no unexplained conclusions to defend in an exam.

§

The audit trail is the product

Evidence is generated while the work happens. The pack an examiner reviews comes from the same pipeline that did the work.

The exam, before the exam

Rehearse the examination before a regulator runs it.

Two features exist for the moments compliance leaders actually lose sleep over: the enforcement action against a peer, and the first-day letter announcing your own exam. RegTwin lets you run both in advance.

Exam simulatorTriggered by peer citations

When a regulator cites someone else, test yourself.

When an enforcement action or MRA against a peer institution is published, the simulator runs the same scrutiny against your framework: which of your controls would have been in scope, where your evidence is thin, and which findings would likely follow at your institution. You see the exposure while there is still time to close it.

OUTPUT: MRA/MRIA RISK VIEW + EVIDENCE CHECKLIST

Document examRun before your exam

Let your own system take the first hard look.

Before an examination, the challenge agent reviews your policies, procedures, and evidence packs the way an examiner would: request by request, document by document. It produces an internal findings list with the gaps ranked, so the first uncomfortable read of your documentation comes from your own system instead of the regulator.

OUTPUT: RANKED INTERNAL FINDINGS + REMEDIATION QUEUE

A note on claims: the simulator flags risk, it does not predict what an examiner will conclude. Anyone who tells you otherwise is selling you an exam finding of your own.

Who built this

Twenty years inside the banks this is built for.

I'm Richard Leclezio. I've spent two decades running change and governance programs at the institutions below, sitting through the exams and owning the remediation plans. I built RegTwin because regulatory change management stayed manual while everything around it modernized, and I got tired of watching that.

SR 11-7BCBS 239MODEL RISK GOVERNANCEREG CHANGE MGMT
Mizuho Financial GroupEnterprise AI
Deutsche BankChange & Governance
Credit SuisseChange & Governance
CitiChange & Governance
Bank of AmericaChange & Governance

PLACES I'VE WORKED. NOT CUSTOMERS.

Founding design partners

I'm looking for three banks to build v1 with.

RegTwin is early, and I'd rather say so than pretend otherwise. If you join now, the twin gets built around your obligation register, your control taxonomy, and your regulators, and you'll have a direct line to me while that happens.

Book a working session

45 MINUTES · BRING ONE RECENT ISSUANCE AND WE MAP IT THROUGH YOUR FRAMEWORK, LIVE · NO SLIDES

Partner terms
§

Your regulators, modeled first

The twin is built against your actual obligation register and agency footprint rather than a generic corpus.

§

Weekly working sessions with me

A standing call, a direct channel, and real influence on what gets built next.

§

Founding pricing, locked

Partner pricing holds for the life of the relationship. That's the trade for helping me get it right.

§

Sandbox first, if you prefer

You can run real issuances through a reference twin on public regulations before any procurement conversation starts.

§

Reference rights stay with you

I'll only use your name as a reference if you agree to it, anonymized or named, in whatever form you're comfortable with.

How it's built

Designed for the diligence questionnaire.

A.1Your data stays yours

Customer documents and obligation data are not used to train models. Retention and deletion terms are written into the contract.

A.2Human approval is enforced in code

The approval gate is enforced in the pipeline itself. There is no bypass flag.

A.3Deploy where your risk team allows

Cloud, virtual private cloud, or on-premises for institutions whose policies require it. SSO and role-based access throughout.

The bet

The next exam cycle is coming either way.

The question is whether your team will still be processing issuances by hand when it arrives. Bring one recent issuance to the working session and we'll run it through end to end.