[ Regulatory AI ]

Your rules,
in the right order.

We turn your organisation's own layered rulebook into an assistant that gives cited, context-specific answers in seconds — respecting which local rule overrides which national one — and never answers beyond what's actually written.

$1M/day

CFPB civil money penalty for knowing violations — the cost of one wrong answer applied consistently across a team

CFPB Regulation X · 12 CFR § 1024.41

$156,259

OSHA willful violation per occurrence — one incorrect safety ruling on a federal worksite

OSHA 29 CFR 1926 · Construction Standard

< 30 sec

Cited, context-specific answer — exact clause, document hierarchy, and what it overrode

[ The problem ]

Most AI tools flatten layered rules into mush.

Regulated bodies live with layered rules: a national standard, a state variation, a local exception, each overriding the one beneath it. Flatten them and you get confident, wrong answers.

Problem 1

The hierarchy gets ignored

Generic AI tools treat all rules as equal. Local exceptions and national defaults get mixed together. The answer might be technically correct for some context — just not yours.

Problem 2

Answers aren't cited

You can't verify what you can't trace. When there's no clause reference attached to an answer, there's no way to know if the ruling is right — or where it came from.

Problem 3

Context-blind answers

The same question can have different answers depending on grade, format, or jurisdiction. A tool that doesn't know which context it's answering for is guessing — and won't tell you it is.

The difference

We model the hierarchy, clause by clause.

The answer you get is the rule that actually governs your situation — not the closest match in a flattened corpus. And you can see exactly what it overrode to get there.

[ What makes it different ]

Not a chatbot. Not document search. A compliance architecture.

The market is full of tools that let you ask questions about your documents. Most are a thin AI layer over basic keyword search. They don't understand document authority, can't resolve a conflict between two policies, and have no memory of how a question has been answered before.

01 — Rule Hierarchy

Understands which document overrides which

The assistant knows the priority of every document in your corpus: legislation sits above internal policy, policy above procedure, procedure above guidance. When rules conflict it doesn't surface the conflict and stop — it resolves it, citing which document governs and why.

02 — Citation Integrity

Every answer cites the exact clause, version, and document

Not a link to a document. Not "see Policy X." The exact section reference: "Section 4.2.1(b) of the Credit Compliance Framework, version June 2025." Every answer is auditable. Every decision is traceable to its source.

03 — Precedent Layer

Prior rulings surface alongside every answer

Every interpretation decision is logged. When the same question comes up again — this year, next year, after staff turnover — the prior ruling surfaces alongside the live rule. Institutional knowledge stays with the organisation, not with the person who left.

04 — Escalation Workflow

A documented path for the calls that genuinely need a human

When a question exceeds the tool's authority, one tap sends a formal referral with question, AI interpretation, cited clause, and context pre-loaded. The escalation is on record. The human who decides gets everything they need before the conversation starts.

Most rulebooks aren't flat — and most AI tools pretend they are. Regulated bodies live with layered rules: a national standard, a state variation, a local exception, each overriding the one beneath it. Flatten them and you get confident, wrong answers. We model the precedence clause by clause so the answer you get is the rule that actually governs your situation.

Layer 3 — Wins

Local Rules

By-laws, chapter rules, local exceptions — override everything below

Layer 2

State / Chapter Layer

Regional conditions — override national defaults

Layer 1 — Baseline

National Standard

Apply where no state or local rule overrides

Example: CFPB regulations → internal credit policy → department procedure

[ How it works ]

Four steps from question to cited answer.

No document codes. No legal training. Ask in plain English — get back the rule that actually governs your situation, cited to the exact clause.

[ 01 ]

Select your context

Choose the relevant policy area, department, or jurisdiction. This scopes every answer to the exact rules that apply — the assistant knows which documents govern which situation.

[ 02 ]

Ask

Type the question in plain English, as you'd ask a knowledgeable colleague. No document codes, no legal terminology, no training required.

[ 03 ]

Get a cited answer

The assistant explains the relevant rule, cites the exact clause and document version, and surfaces any prior interpretations on the same question. Auditable from the moment it's given.

[ 04 ]

Escalate if needed

When the situation genuinely needs a human call, one tap sends a formal referral with full context pre-loaded. The existing process stays intact — this removes the noise so real decisions get proper attention.

AI

Regulatory AI Assistant

Powered by your organisation's documents

Online
Question — Financial Services
A borrower mentioned financial difficulty during our serviceability call. What are our obligations from here?
Assistant
Once a borrower mentions financial difficulty, two obligations trigger immediately. First, offer to discuss hardship assistance options under your Hardship Policy — this cannot be deferred. Second, log the disclosure in your CRM within 24 hours with the date, channel, and content of the conversation. If a hardship arrangement is requested, the 21-day IDR response clock starts from that request. Reference · Internal Hardship Policy v3.2, Section 5.1 · CFPB Regulation X, 12 CFR § 1024.41 — Loss Mitigation Procedures (overrides internal policy where more stringent)
[ Who it's for ]

Any organisation where rules are documented but the people applying them aren't compliance experts.

The same architecture works across industries because the problem is the same: layered documents, frontline staff applying them under pressure, and a compliance team that becomes the bottleneck.

Financial Services

CCO · Chief Compliance Officer · Registered Investment Adviser

Loan officers, advisers, and brokers interpret overlapping federal and state rules at the point of sale. The compliance team learns about misapplication when the CFPB or state regulator is already involved. Enforcement appetite across lending, disclosure, and fair lending has intensified — UDAAP violations and redlining patterns are among priority targets.

CFPB civil money penalties: up to $1M per day for knowing violations. A single UDAAP pattern across 500 loans: $2M+ remediation exposure plus restitution orders.

Healthcare & Aged Care

Director of Nursing · Chief Quality Officer · Medicare/Medicaid Provider CEO

25–35% staff turnover annually means the interpretation problem resets every season. Directors of Nursing field protocol questions while running the floor. CMS billing rules and coverage criteria update frequently — claiming errors are common and expensive. Survey visits surface misapplication that has been accumulating for years.

CMS false claims: $13,000–$26,000 per claim under the False Claims Act. OSHA willful violation: up to $156,259. CMS decertification halts all Medicare/Medicaid reimbursement.

Construction & Safety

National Safety Manager · HSE Manager · Site Manager

Distributed workforce, high subcontractor turnover, and licensing questions that are jurisdiction-specific. Updated OSHA silica and heat illness rules changed obligations across the sector and most businesses are still catching up. Safety officers spend significant time on "can this person do this work" questions that should be self-serve.

OSHA willful violation: up to $156,259 per violation. A single recordable fatality: $1M+ average total cost. Licensing non-compliance on a federal or state tender can void the contract.

Professional Associations

CEO · Director of Member Services · Registrar

Member services teams field thousands of "do I comply?" questions by phone and email. The wrong answer means a member practices with a lapsed licence, fails their CPD requirement, or breaches a condition they didn't know existed. The member carries the liability — the association carries the reputational risk.

Practising without a valid licence: criminal offence in most US professions. A state board investigation for a member: $30–150K in legal fees, public disciplinary record, income loss.

[ What your organisation gets ]

For your leadership and frontline.

Your compliance team fields fewer repeat questions. Every decision is traceable to its source. And when a regulator asks, you have an auditable record of what rule was applied, from which document, and when.

01

One source of truth across the entire organisation

The same question gets the same answer regardless of who's asking, which site they're on, or what time it is. Policy inconsistency — and the liability it creates — stops being a problem.

02

Your compliance bottleneck disappears

Routine questions are resolved by the tool. Your compliance team focuses on the judgement calls that genuinely need them, not the same three questions fielded for the fifth time this week.

03

Every escalation arrives complete

When a matter genuinely needs a human call, it comes through as a full referral: question, AI interpretation, cited clause, and context already captured. No phone tag. No re-explaining. The decision can be made immediately.

04

Audit trail for every policy decision

Every question is logged. Every referral is documented. When a regulator, auditor, or lawyer asks what was applied and when, the record is immediate, complete, and requires no reconstruction.

05

Institutional knowledge that survives staff turnover

Prior interpretations and rulings are logged and surface alongside future answers. Compliance knowledge stays with the organisation, not with the person who leaves at the end of the year.

06

See the gaps in your policies before they become breaches

Questions your documents don't clearly answer become visible. Most organisations discover genuine ambiguities in policies they thought were clear — you can fix them before they produce an inconsistent decision or a regulatory finding.

[ Investment ]

Unlimited users. Priced by organisation size, not headcount.

Per-seat pricing punishes adoption. Every team member should be using this. The annual licence covers unlimited users — the rate scales with your organisation size so upgrading only happens when it actually saves you money.

Starter

Starter

$8,500

+ applicable state tax / year

Up to 100 employees · unlimited users

+$45 per employee over 100

Setup & onboarding: $3,000 · once off
Scale

Scale

$22,000

+ applicable state tax / year

Up to 500 employees · unlimited users

+$30 per employee over 500

Setup & onboarding: $8,000 · once off

All tiers include cited Q&A, escalation workflow, audit log, admin dashboard, quarterly policy gap report, and ongoing document updates. Growth and Scale add custom branding and HRIS/intranet integration options. Setup covers document ingestion, rule hierarchy mapping, citation layer build and testing, and initial policy gap analysis — handled entirely by us, nothing for your team to configure or maintain. Enterprise (500+ employees with complex needs): contact us directly.

[ Proof of concept ]

In development: Cricket Tasmania.

The first live deployment of this engine. A signed engagement — currently being built.

Active build

Laws of Cricket AI Assistant

Cricket Tasmania manages 507 senior teams and 8,750 registered players across 150+ clubs. Time-poor volunteers were fielding constant rules questions they didn't always know the answers to — pulling from the MCC Laws of Cricket, CT's playing conditions, and CT's own by-laws simultaneously. We're building an assistant that resolves that hierarchy on every question, cites the exact clause, and delivers a context-specific answer in seconds.

507 senior teams
8,750 registered players
3 layers MCC Laws → CT playing conditions → CT by-laws — resolved clause-by-clause
Read the build log →
[ Get started ]

Have a rulebook that needs an assistant?

Tell us about your corpus — the documents, the layers, and the people who need to interpret them. We'll tell you honestly whether this is the right fit and what a build would look like.

Got it.

We'll be in touch. — Bailey