FINANCIAL CRIME COMPLIANCE · FCA-AUTHORISED FINTECHS · UK · AML · KYC/UBO · SANCTIONS SCREENING · SAR DRAFTING · EVIDENCE PACKS

AI agents for financial crime compliance
— the MLRO always signs off.

Financial crime compliance is the obligation that carries personal, criminal liability for the MLRO at an FCA-authorised fintech. VIRGIL is the AI-assisted financial crime layer — investigating transaction monitoring alerts, screening sanctions, PEP and UBO exposure, and drafting SAR narratives with full evidence provenance. VIRGIL prepares, drafts and evidences; your team reviews and signs off. Nothing is filed without a human.

FCA enforcement actions are up 43% in the past 24 months, and alert volumes keep rising against compliance teams of one to three. VIRGIL investigates the alert, evidences the decision and drafts the SAR — every conclusion traced to its source transaction, so the record survives an FCA s166 review.

Platform Features

How a compliance obligation
gets resolved in 48 hours.

VIRGIL connects to your transaction data, runs a six-agent consensus engine built on documented AML and financial crime failure patterns, and drafts SARs, alert rationales and evidence packs mapped to the UK financial crime stack — POCA 2002, MLR 2017, JMLSG guidance, FCA SYSC 6.3, the OFSI sanctions regime and the NCA SAR/DAML regime — so your team reviews and submits, never drafts from scratch.

F-001
Compliance automated. No manual assembly required.

VIRGIL monitors your transaction stream and infrastructure posture, scores anomalies, and generates regulatory reports automatically. Your compliance team focuses on decisions and sign-off — not pulling data, writing SARs, or formatting audit packs.

DATA → SCORE → REPORT → SIGN-OFF
F-002
High-accuracy AML detection — built on documented failure signatures

The agent engine is built on documented AML and financial crime failure patterns — the specific pre-event signals in transaction velocity, structuring behaviour, and counterparty patterns. Six agents vote in consensus before any alert is escalated, filtering low-risk alerts for human confirmation and cutting false-positive review time.

AML FAILURE SIGNATURES · 6-AGENT CONSENSUS
F-003
SAR drafts in one click. FCA-ready.

Compliance analysts spend 60–70% of their week manually drafting SARs. VIRGIL Copilot generates complete Suspicious Activity Report drafts mapped to FCA requirements, ready for MLRO review and submission — in a single click.

ONE-CLICK SAR · FCA-MAPPED · MLRO READY
F-004
Your UK financial crime obligations, evidenced continuously

POCA 2002, MLR 2017, JMLSG guidance, FCA SYSC 6.3, the OFSI sanctions regime and the NCA SAR/DAML regime — the UK financial crime stack your fintech answers for, evidenced as you work each alert. VIRGIL keeps the record a supervisor can read.

POCA 2002 · MLR 2017 · JMLSG · SYSC 6.3 · OFSI
F-005
Audit evidence packs — ready for your next supervisory visit

VIRGIL generates timestamped, event-level evidence bundles mapped to the UK financial crime stack — a decision trail per alert and SAR, structured for FCA s166 reviews, REP-CRIM support and sponsor-bank reporting.

AUDITABLE · s166-READY · REP-CRIM
F-006
Free gap report. Live in under 48 hours.

Read-only API connection. 48 hours. VIRGIL delivers a detailed report showing your top 5 financial crime blind spots against your UK obligations — MLR 2017 monitoring, sanctions screening, SAR readiness and evidence gaps. No sales call required.

FREE GAP REPORT · READ-ONLY · ZERO DISRUPTION
Inside the platform
The Command Center your
compliance team works in.

Live readiness scores, AI-ranked actions, a regulatory-intelligence feed, a risk heatmap, evidence coverage, AML anomaly detection and one-click SAR drafts — every obligation tracked in one workspace.

Impressive — you didn't just bolt an LLM on a problem. You understood where AI shouldn't be trusted and forced every SAR draft through human review. That's the judgment that separates compliance experts from armchair theorists.

VP
Vatsal Pareshkumar
Founder · OutputLens
VIRGIL Copilot — Now Live

Your compliance team's
always-on co-pilot.

VIRGIL Copilot is the MLRO-facing intelligence layer — surfacing agent reasoning, flagging live alerts, and generating SAR drafts directly inside your compliance workflow. No tab-switching. No copy-pasting from a dashboard. Compliance intelligence where your team already works.

VIRGIL Copilot — a live FCA-compliant SAR draft with the six-agent reasoning panel (AML Review, UBO Verifier, Network Analyst, Policy Alignment, Predictive Forecast, Format & NCA Compliance)
SAR Draft in One Click

Copilot generates a complete, FCA-mapped Suspicious Activity Report draft from the flagged transaction data — with agent reasoning, counterparty analysis, and structured sections ready for MLRO review. One click. Under 60 seconds.

Agent Reasoning Visible

Every alert shows the full 6-agent consensus trace: which agents fired, what signals triggered them, and why. MLROs sign off on evidence, not black boxes — satisfying FCA explainability expectations out of the box.

Live Alert Triage

Alerts surface prioritised by severity, framework relevance, and velocity signal. Copilot filters low-risk alerts for human confirmation before they reach the MLRO queue — cutting false-positive review time while raising escalation quality.

Sanctions, PEP & UBO Screening

Copilot cross-references counterparties against sanctions and PEP lists and flags beneficial ownership not yet verified against Companies House — surfacing the enhanced-due-diligence gaps an MLRO must close before filing.

See Copilot in action

A 30-minute demo includes a live Copilot walkthrough on your specific regulatory obligations.

Book a Demo
Built For

Built for FCA-authorised
fintechs.

VIRGIL's focus is FCA-authorised fintechs with 1–3 person compliance teams — where a missed SAR, a failed audit, or a compliance gap surfaced in diligence carries deal-blocking, licence-risk consequences.

PRIMARY ICP
01
MLROs & CCOs at Series A–B Fintechs

Our beachhead. MLROs at FCA-authorised fintechs face personal liability under FCA MLRO Rules and supervisory visits rising 40% year-on-year. VIRGIL gives them a live, auditable compliance record — AML anomaly detection, SAR drafts, and evidence packs — that turns an overwhelmed 2-person team into a defensible compliance function.

UK Beachhead · FCA-Authorised · EMI · BaaS
02
CFOs & COOs at Series B Pre-Close

Compliance gaps surfaced during Series B due diligence kill deals. VIRGIL generates the audit trail and evidence packs that satisfy investor DD teams before they flag compliance as a deal risk. The trigger: a Series B round in the next 6 months with DD starting.

Deal-Risk Evidence
03
BaaS Operators & Sponsor Banks

Sponsor banks tightening quarterly reporting requirements create acute compliance pressure for BaaS operators. VIRGIL generates sponsor bank report templates automatically and surfaces the operational risk dashboard that satisfies quarterly reporting without manual effort.

Sponsor Bank Reporting
04
Crypto / VASP Firms Under FCA Registration Pressure

Firms in or approaching FCA cryptoasset registration face intense scrutiny of their AML systems and controls. VIRGIL supplies the investigated alerts, sanctions/UBO screening logs, and evidenced trail that a registration or supervisory review expects — fast.

VASP · MLR 2017 Registration
Why VIRGIL Wins

The moat is the
regulatory pattern library.

VIRGIL's agent engine is built on documented AML and regulatory failure patterns — the specific pre-event signals that precede regulatory action. That pattern library is the moat.

Read-Only, Evidence-First Architecture

VIRGIL connects by read-only API and never writes to your systems. Every conclusion carries its provenance — source transaction, rule triggered, agent reasoning, human sign-off — built for an FCA s166 review from day one.

End–to–End Encryption

Data encrypted in transit (TLS) and at rest (AES–256). Zero–plaintext policy for production systems handling transaction, KYC, infrastructure, and regulatory data.

Resilient Infrastructure

Automatic failover and continuous availability, with UK data residency available for FCA-authorised clients with jurisdiction-specific requirements.

Enterprise SLAs & DPA

Contractual SLAs, data processing addenda, and security exhibits aligned with FCA-authorised firm, BaaS operator, and enterprise buyer requirements.

Regulatory Coverage

The UK financial
crime stack.

VIRGIL covers your UK entity's financial crime obligations under UK law — POCA 2002, MLR 2017, JMLSG guidance, FCA SYSC 6.3, the SMCR (SMF17), OFSI sanctions and the NCA SAR/DAML regime, with REP-CRIM data support. It does not claim to cover other regulators' regimes. Continuously evidenced, with the trail to prove it.

R-001
POCA 2002 · Active

The Proceeds of Crime Act s.330 disclosure obligation — failure to disclose is a criminal offence. VIRGIL drafts the SAR narratives and evidences the disclosure decision for the MLRO to review and submit.

ACTIVE · s.330 DISCLOSURE
R-002
MLR 2017 · Active

The Money Laundering Regulations: risk-based ongoing monitoring, CDD and EDD. VIRGIL investigates alerts and produces the monitoring evidence and CDD/EDD rationale a supervisor expects to see.

ACTIVE · RISK-BASED MONITORING
R-003
JMLSG Guidance · Active

Joint Money Laundering Steering Group good practice. VIRGIL aligns its CDD/EDD rationale and counterparty write-ups to JMLSG, so the reasoning behind each disposition is defensible.

ACTIVE · CDD / EDD RATIONALE
R-004
FCA SYSC 6.3 · Active

Systems and controls against financial crime, and SMCR (SMF17) accountability. VIRGIL evidences that monitoring ran, that alerts were worked, and that decisions were recorded — the systems evidence SYSC expects.

ACTIVE · SYSC 6.3 · SMF17
R-005
OFSI Sanctions Regime · Active

Financial sanctions compliance. VIRGIL screens counterparties against sanctions and PEP exposure, flags UBO gaps against Companies House, and logs every check with a timestamp.

ACTIVE · SANCTIONS · PEP · UBO
R-006
NCA SAR / DAML · Active

Suspicious Activity Reports and Defence Against Money Laundering. VIRGIL drafts SARs and DAML requests for human review; the MLRO submits to the NCA. VIRGIL never files.

ACTIVE · SAR / DAML DRAFTING
FAQ

Frequently asked
questions.

Screening tools detect — they flag a name or a transaction and hand you an alert. VIRGIL investigates and evidences: it works the alert through a six-agent consensus review, writes the rationale for the disposition, drafts the SAR and keeps a timestamped trail for your s166 file. VIRGIL does financial crime only — no SOC 2, no operational risk, no security evidence. That focus is deliberate: a tool that tries to do every framework does none at the depth an FCA supervisor tests. It runs alongside your existing monitoring, and it's priced for Series A–B fintechs rather than banks with six-figure budgets.

VIRGIL Copilot is the MLRO-facing intelligence layer built into the platform. It surfaces live alerts with full agent reasoning traces, generates SAR drafts in one click, and exports evidence packs — all inside the compliance workflow. See it demonstrated at any product demo.

No. VIRGIL automates the document assembly, report generation, and monitoring that currently consumes 60–70% of a compliance analyst's week. Your MLRO reviews SAR drafts via Copilot and submits them — they never draft from scratch. FCA MLRO Rules require personal sign-off; VIRGIL makes that sign-off faster and better-evidenced, not unnecessary.

Connect a read-only API, or upload your transaction data as CSV, Excel, or PDF. Within 48 hours, VIRGIL delivers a detailed report showing your top 5 financial crime blind spots against your UK obligations — MLR 2017 monitoring, sanctions screening, SAR readiness and evidence gaps. No sales call required, and no commitment attached.

VIRGIL connects by read-only API and never writes to your systems. Data is encrypted in transit (TLS) and at rest (AES–256), with least–privilege access controls, network segmentation and continuous monitoring. UK data residency is available for clients with jurisdiction-specific requirements. A Data Processing Addendum is available on request.

Book a Demo

See VIRGIL in action.

Book a 30-minute demo and we'll walk you through VIRGIL's AML detection, SAR generation via Copilot, and audit evidence capabilities using a live environment. Our team will respond within 24 hours.

WHAT TO EXPECT
Live Platform Walkthrough
See AML anomaly detection, Copilot SAR generation, and audit evidence packs on a live environment.
Your Use Case
We tailor the demo to your obligations — POCA 2002, MLR 2017, OFSI sanctions, SAR drafting, or sponsor-bank AML reporting.
Free Gap Report
Every demo includes a complimentary AML or infrastructure blind spot assessment — no commitment required.
30 Minutes. No Sales Pitch.
A focused technical walkthrough with our compliance engineering team. You leave with a clear view of how VIRGIL fits your stack.
Current Availability

Accepting demo requests for Q2 2026. Response within 24 hours.

DEMO REQUEST FORM

By submitting you agree to our Privacy Policy. 7EulerianStrategies.

P(SAR|D) ∝ P(D|SAR)·P(SAR) E[fine] = Σ p¹·F¹·(1−m) precision = TP / (TP + FP) risk = f(velocity, θ, counterparty) AML ∩ KYC ∩ OFSI
7EULERIANSTRATEGIES · ABOUT US FOUNDED 2025 · UK
ABOUT / 7EULERIANSTRATEGIES

Engineering
the compliance automation
era.

We automate regulatory compliance for UK fintechs before it becomes a crisis. VIRGIL continuously monitors transactions, detects AML anomalies, generates SARs via Copilot, and produces audit-ready evidence packs. Our regulatory pattern library is the moat.

The industry doesn't have a data problem. It has an automation problem — compliance signals arrive too late, too manually-processed, and too siloed to prevent regulatory action. We built VIRGIL to close that loop.

Kent & Kingston
University roots
Founded by engineering and research undergraduates united by a shared belief that fintech compliance deserves better tools and better outcomes.
Production–Grade
Engineering culture
Large–scale agentic AI systems, production data pipelines, and real–world regulatory exposure — from day one.
UK · EU · Global
Global ambition
Building the compliance automation layer for the world's most regulated fintech and infrastructure markets.
Our Story

Regulatory compliance.
One pattern library.
One moat.

Every major regulatory failure follows a pattern. A suspicious transaction isn't flagged without warning signals — velocity spikes, structuring behaviour, counterparty anomalies. A compliance gap doesn't materialise overnight. The signals are there. The industry just wasn't processing them fast enough, or automatically enough, to act before the FCA did.

We documented AML and regulatory failure patterns — the specific pre-event signal sequences that precede regulatory action, failed audits, and licence revocations. Not from a generic dataset. From direct research and case-by-case analysis, extracting the compliance patterns that standard rule-based monitoring tools consistently miss.

That corpus is encoded into VIRGIL's agent engine and surfaced through Copilot. It is the reason VIRGIL catches compliance risks before they escalate.

We then wired that engine to transaction data streams, infrastructure monitoring, regulatory framework mappings, and SAR generation workflows. The result: a platform that detects AML anomalies, scores severity, generates regulatory reports, and compiles audit evidence — automatically, in under 48 hours.

WHAT WE STAND FOR
  • Engineering rigour
    Production–grade systems, not prototypes. We ship compliance automation that holds under real regulatory conditions.
  • Regulatory depth
    We understand POCA 2002, MLR 2017, JMLSG, FCA SYSC 6.3 and the SMCR from the inside. We know what an MLRO needs to sign off a SAR.
  • Automate before, not remediate after
    The industry is built on reactive compliance. We're building the infrastructure to shift it towards prevention.
  • Team over hierarchy
    We move fast because we trust each other. Decisions are made by people closest to the problem.
  • Craft meets impact
    We genuinely care about what we build — the accuracy of the agents, the precision of the SAR drafts, the clarity of the audit packs.
Founder

Mitual Sharma.
Founder, CEO & CTO.

Building AI-native compliance infrastructure for the next generation of fintechs.

Most compliance failures are not surprises. The signals exist — in transaction velocity, counterparty behaviour, control gaps — long before a regulator acts. The industry's problem was never data. It was that the data arrived too late, too manual, and too disconnected to prevent the outcome.

His edge is the combination this market rarely holds in one person: AI systems engineering, regulatory operating knowledge, and product architecture. He builds the detection models, designs the agent consensus that makes their output trustworthy, and turns dense regulatory obligation into something a founder can act on in an afternoon — not a quarter.

Compliance should be invisible, automated, and on the founder's side — not a tax on building.

He started by building risk-prevention systems and kept hitting the same wall: compliance teams of one or two, expected to hold back obligations designed for departments. VIRGIL is the tool he wished those teams had — one that does the watching, drafting, and evidencing, so people can do the judgement.

VIRGIL is the financial crime engine for FCA-authorised fintechs — investigating, evidencing and drafting, so the MLRO can do the judgement.

WHAT HE LEADS
  • AI systems design
    The detection models and six-agent consensus that make AI-assisted compliance output something a team will trust and act on.
  • Product architecture
    From a transaction or infrastructure signal to an MLRO-ready SAR, audit pack, and live readiness score.
  • Regulatory intelligence
    Models that keep VIRGIL current as FCA, MLR 2017, sanctions and SAR obligations move — and tailored to each firm.
  • Commercial narrative
    How the platform works — and why it matters to the people who carry the regulatory risk.
The Team

The team that built
the pattern library.

VIRGIL's moat is a documented corpus of AML and regulatory failure patterns. Building it required AI engineers who understood anomaly detection at production scale, platform engineers who could wire transaction signals to real-world regulatory reports, and domain experts who knew what those signals actually meant inside an FCA-authorised firm's compliance function.

MS
Mitual Sharma
Founder · CEO & CTO

Owns VIRGIL's technical direction and commercial narrative — the detection models, the 6-agent consensus, the product architecture, and the regulatory intelligence that keeps it current. AI engineering, regulatory operating knowledge, and product in one place.

Agentic AIProduct ArchitectureRegulatory Intelligence
AM
Alex Magena
Growth & Operations

Brings real operating experience from building and running earlier ventures — growth, brand, and commercial execution. The muscle that turns a precise compliance product into a business fintech founders and their banks actually buy.

Go-To-MarketOperationsCommercial
TS
Teghvir Singh
Co-Founder · Engineering & Systems

Behind the production systems that make VIRGIL real — the real-time pipeline from signal to SAR, the framework mappings across POCA 2002, MLR 2017, JMLSG and the UK financial crime stack, and the infrastructure that holds under genuine regulatory conditions, at scale.

Real–Time SystemsData InfrastructurePlatform Engineering
Operating track record — ventures the team has built & run: AllIsWellOnline MarketingZend Rochards
Connect With Us Follow our progress at 7EulerianStrategies
Our Mission

Rigorous engineering.
Deep regulatory
understanding.

We exist to create AI that genuinely serves the people solving real–world compliance problems — MLROs, CCOs, compliance officers, and analysts buried in manual reporting obligations. Not another dashboard. An intelligence that acts, learns, and makes the regulatory function stronger every time it generates a report.

HOW WE'RE DOING IT
  • Agentic AI that closes the compliance loop
    Complete, evidenced regulatory outputs drafted in under 48 hours, from transaction signal to MLRO-ready SAR via Copilot — always for a human to review and submit.
  • Six agents working in consensus, not single bets
    Fewer false positives, stronger recall, and explainability that compliance teams and FCA supervisors can trust and evidence.
  • Built into existing compliance workflows
    We don't ask MLROs to change tools. VIRGIL Copilot surfaces intelligence inside the systems they already use every day.
  • Enterprise–grade from the start
    Read-only by default, end–to–end encrypted, UK data residency available. The security posture FCA supervisors expect — VIRGIL never writes to your systems and never files.
Work With Us

See what VIRGIL can do
for your compliance function.
Book a demo.

If you're a UK FCA-authorised fintech with a compliance obligation in the next 6 months — book a 30-minute demo and we'll show you exactly how VIRGIL and Copilot address your specific requirements.

RESPONSE WITHIN 24 HOURS

Pricing

Compliance that scales
with your firm.

Subscription tiers scaled to your stage and alert volume. Every tier includes the six-agent investigation engine, SAR Copilot, sanctions/PEP/UBO screening and exportable evidence packs — covering your UK financial crime obligations. Every tier keeps the MLRO in command: VIRGIL drafts and evidences; your team reviews and signs.

Available now
Startup
£2,000/month
For pre-seed and seed FCA-authorised fintechs standing up financial crime compliance with a team of one to three.
  • AI Financial Crime Copilot — SAR drafting
  • Six-agent alert investigation
  • Sanctions, PEP & UBO screening
  • FCA, AML & SAR readiness scores
  • Live regulatory updates — FCA, NCA, OFSI
  • Exportable evidence packs — s166-ready
  • Read-only transaction data connection
  • Unlimited API access · Up to 5 users
Covers your UK financial crime stack — POCA 2002, MLR 2017, OFSI, SAR/DAML
Book a Demo
Available now
Enterprise
Custom pricing
For larger FCA-regulated firms needing contractual SLAs, a DPA and dedicated support across multiple UK entities.
  • Everything in Growth, plus…
  • REP-CRIM return data support
  • Multi-entity UK financial crime management
  • Enterprise governance & advanced approval chains
  • Contractual SLAs, DPA & UK data residency
  • Dedicated onboarding · SLA-backed support
  • Unlimited users
Contractual SLAs, DPA and UK data residency for regulated firms
Book a Demo

VIRGIL covers your UK entity's financial crime obligations. It prepares, drafts and evidences — it does not certify compliance or file with the NCA. Those decisions and filings remain with the firm.

P(SAR|D) ∝ P(D|SAR)·P(SAR) E[fine] = Σ p¹·F¹ precision = TP / (TP + FP) risk = f(velocity, θ) AML ∩ KYC ∩ POCA
LIVE RESEARCH HUB VOLUME IV · Q1 2026 7EULERIANSTRATEGIES · INSIGHTS
002 / REGULATORY INTELLIGENCE INSIGHTS

Behavioural
science meets
fintech compliance.

Academic research, practitioner analysis, and evidence-based perspectives on how regulatory frameworks, cognitive biases, and AI automation shape compliance decisions — from AML judgment to audit readiness and regulatory risk perception.

24Articles Published
4Research Streams
Q1 2026Latest Edition
Regulatory Behaviour
Stream 01 · 6 articles
Heuristics, biases, and cognitive shortcuts that distort compliance risk perception in fintech teams and MLROs under regulatory pressure.
AML Decision-Making
Stream 02 · 6 articles
Dual-process theory, framing effects, and groupthink in AML alert triage and suspicious activity judgments.
Enforcement & Psychology
Stream 03 · 6 articles
Loss aversion, prospect theory, and the endowment effect applied to FCA enforcement response and regulatory settlement negotiation.
AI & Compliance Futures
Stream 04 · 6 articles
Temporal discounting, automation bias, and trust calibration in AI-assisted compliance and agentic regulatory systems.
FEATURED / MOST RECENT
SHOWING 8 OF 23 ARTICLES
Interactive Research Tools

The mathematics of
compliance risk.

Four foundational models from decision science and regulatory economics, rendered interactively.

01 · Bayesian AML Updating
P(SAR|D) = P(D|SAR) · P(SAR) / P(D)
Prior P(SAR)0.30
Alert accuracy0.75
False +ve rate0.15
Posterior P(SAR|alert): —
Updating prior SAR probability with new transaction anomaly evidence. Critical for understanding how real-time AML signals should shift an MLRO's prior on expected suspicious activity.
02 · Prospect Theory in Compliance
v(x) = x¹ (gains) · −λ(−x)² (losses)
Loss aversion λ2.25
Gain curve α0.88
Loss aversion ratio: 2.25× gains
Kahneman & Tversky's value function explains why MLROs feel a regulatory fine far more acutely than an equivalent compliance investment saving.
03 · Expected Regulatory Loss
E[L] = Σ¹ p¹ · F¹ · (1 − m¹)
P(AML fine)8%
P(licence action)3%
Automation m35%
Expected annual regulatory loss: —
Computes expected regulatory loss across risk categories with compliance automation mitigation applied.
04 · AML Signal Detection Theory
d′ = (μ_signal − μ_noise) / σ · β = threshold
Sensitivity d′2.0
Threshold β0.0
Hit rate: — · False +ve: —
Models the tradeoff between missed suspicious transactions and false AML alerts in AI-assisted compliance monitoring.
Category · Date

Article Title