Lýseis Verity — Field Manual · v2.0

A guide to running compliance on Verity.

This manual explains what Verity does, how each of its 13 operational domains works, and who on your team touches which parts of it — written for new staff learning the system and for institutions evaluating it.

SOURCE → SIGNAL → INTELLIGENCE → DECISION

If you're onboarding

Start at Getting Started, then read your role's domains from the sidebar.

If you're evaluating Verity

Read What Verity Is, then Security & Trust for how it protects your data.

If you manage the team

Roles & Permissions covers exactly what each of your six roles can do.

Orientation

What Verity Is

A financial crime and risk intelligence platform — not a decision-maker.

Verity watches the transactions, loan activity, payments, and identity signals flowing through a financial institution, and turns them into risk scores, alerts, and investigative leads. It is built for the teams responsible for AML/CFT compliance — the people who have to know, for every customer, whether something looks wrong, and have to be able to prove to a regulator exactly why they decided what they decided.

Verity's one non-negotiable design rule is in its own name: it tells you what it sees, never what you should do. A risk score, a triggered rule, a watchlist hit — these are signals. The decision to restrict an account, escalate to enhanced due diligence, or file a statutory report is always made by a named, accountable person inside your institution, and Verity records exactly who made it and why.

Source systems (your core banking platform, lending engine, payment gateway, screening provider) push events in. Verity canonicalizes them, scores risk, and runs them against your institution's declarative rules. What fires becomes an alert. What an alert resolves into — a case, a decision, a filing — is where a human takes over.

The 13 domains, at a glance

Every screen in Verity belongs to one of these. Click any to jump to its section.

Orientation

Getting Started

From invitation link to your first signed-in session.

  1. Accept your invitation. A compliance officer or admin at your institution invites you by email. The link opens a page confirming your assigned role — set your password there.
  2. Set up two-factor authentication. Verity requires it for every account, no exceptions. Scan the QR code with an authenticator app (Google Authenticator, Authy, 1Password) and confirm with the 6-digit code it shows.
  3. Sign in. Email, password, then your 6-digit authenticator code. Every session re-checks this — there's no "remember this device."
  4. Land on your workspace. Where you land depends on your institution's status: a brand-new institution goes into the onboarding wizard; an approved one goes straight to the Overview dashboard.
If your institution is new

Onboarding is a five-step wizard — organization details, licensing & regulatory info, KYC document upload, an indemnity agreement, then review. A Verity compliance reviewer approves or requests changes; you'll see the status update live once a decision is made.

Orientation

Roles & Permissions

Six roles. What each one can actually touch.

Role Alerts & Cases Rules & Risk Config Reports Team & Org Settings
Tenant Admin Full Full Full + approve Full
Compliance Manager Full Read only Full + approve —
Investigator Work & resolve Read only Prepare (not approve) —
Auditor Read only Read only Read only —
Analyst (read-only) Read only — — —
Integration — — — —

"Integration" isn't a person — it's the role an API key operates under when a source system (your core banking platform, for instance) pushes events into Verity directly. It can write events and screening results, nothing else.

One rule applies everywhere regardless of role: whoever prepares a regulatory filing can never be the one who approves it. See Regulatory Reports.

01 · Operational Core

Overview

The first screen you see. Built to answer one question: what needs me today?

All roles

Overview is a morning-briefing screen, not a report. It shows your open alert queue sorted by SLA urgency, active case count, and 24-hour event throughput, so the first thing you see when you sign in is whatever actually needs a decision — not a wall of charts.

  • SLA countdowns on every open alert, color-coded by how close it is to breach.
  • Active case count — investigations currently open, not yet decided.
  • 24h throughput — how much is flowing through ingestion, for a sense of normal volume.
02 · Operational Core

Alerts

Where every triggered rule or risk event lands first.

Tenant AdminCompliance ManagerInvestigator

Every time a rule fires or a risk threshold is crossed, an alert appears here — never silently, and never auto-resolved. Each alert carries a severity (CRITICAL down to LOW), a priority, and an SLA clock that starts the moment it's created.

  • Assign an alert to an investigator — it moves from NEW to ASSIGNED.
  • Resolve with a mandatory outcome (false positive, EDD required, escalate, restrict activity) and a note. Verity will not let you close an alert without stating why.
  • Convert to Case when triage isn't enough and the finding needs a formal investigation.
  • Alerts that are deduplicated against a recent identical signal won't flood the queue twice.
Once resolved, it's locked

A resolved or case-converted alert can't be quietly reassigned or reopened from this screen. That's deliberate — it's what makes the queue a reliable record of what was actually decided, not just what's currently pending.

03 · Operational Core

Investigator 360°

Everything Verity knows about one customer, on one screen.

Tenant AdminCompliance ManagerInvestigatorAuditor

When an alert needs a closer look, this is where you go. It's a consolidated workspace, not a set of tabs you have to stitch together yourself — identity verification, current risk score with its full factor breakdown, screening history, and recent transactions, all for one subject.

  • Identity & verification — BVN/NIN traceability, masked by default.
  • Risk score evolution — the trail of how a score moved over time (e.g. 18 → 38 → 57 → 81), not just today's number.
  • 7-factor driver breakdown — exactly which signals contributed how much to the current score. Nothing here is a black box.
  • Screening signals — watchlist or sanctions hits, always labeled as signals, never as a verdict.
  • Recent transactions — the activity that's actually driving the picture.
04 · Operational Core

Cases

Where an alert becomes a formal, accountable investigation.

Tenant AdminCompliance ManagerInvestigator

A case bundles one or more related alerts under a single investigation, with its own notes log and a formal decision record. It's the unit a regulatory filing gets built from.

  • Linked alerts — every alert that fed into this investigation, in one place.
  • Investigator notes — a running, timestamped log of findings, attributed to whoever wrote each one.
  • Formal decisions — NO_ACTION, EDD (enhanced due diligence), RESTRICT, or REPORT_REQUIRED, each with a mandatory rationale.
Decisions are permanent

Once a decision is recorded on a case, there's no edit button. If circumstances change, you record a new decision — the old one stays exactly as it was made. This is what makes a case file defensible to a regulator months later.

05 · Data & Intelligence

Customers

Your verified registry of individuals and businesses.

Tenant AdminCompliance ManagerInvestigatorAuditor

Every customer your source systems have reported on lives here, with PII masked by default (an email shows as e***@example.ng, never in full) regardless of who's looking — internal staff included. This is the registry Investigator 360° and Transactions both draw from.

06 · Data & Intelligence

Transactions

The canonical ledger, normalized across every source system.

Tenant AdminCompliance ManagerInvestigatorAuditor

Whether a transaction originated in your core banking platform, a lending product, or a payment gateway, it lands here in one consistent shape — amount formatted in local currency (₦3,000,000.00), direction, and channel clearly marked, so an investigator never has to guess which system a figure came from.

07 · Data & Intelligence

Risk Engine

A deterministic score, not a model you have to trust blindly.

Tenant AdminCompliance Manager (read)Investigator (read)

Every customer's risk score (0–100) is computed from seven weighted factors — screening, transaction behaviour, identity, lending behaviour, customer profile, payment behaviour, and historical alerts. Tenant admins can tune the weight each factor carries and where the LOW/MEDIUM/HIGH/CRITICAL bands sit, versioned so you can see exactly when a change took effect.

BandScoreTypical recommendation
LOW0–29Standard KYC
MEDIUM30–59Enhanced monitoring
HIGH60–79Review required
CRITICAL80–100Hold & review
08 · Data & Intelligence

Rules

Your institution's detection scenarios, declared, not coded.

Tenant AdminCompliance Manager (read)Investigator (read)

A rule is a structured, declarative condition — not a script, and never arbitrary code — covering thresholds, velocity, frequency, aggregation, behavioural deviation, dormancy reactivation, and structuring (repeated transfers kept just under a reporting threshold). Every rule that fires creates an alert with the severity you configured.

  • Dry-run testing — run a rule against real historical records before it ever creates a live alert, so you know what it would have caught.
  • Versioned — every change to a rule's condition creates a new version; nothing is silently overwritten.
  • Enable / disable independently of editing, so a rule can be paused without losing its configuration.
09 · Data & Intelligence

Monitoring

Behavioral baselines — is this customer still acting like themselves?

Tenant AdminCompliance ManagerInvestigator

Monitoring tracks each customer's rolling baseline — typical monthly volume, typical transaction size — and flags the gap when current activity departs from it sharply. A customer whose baseline is ₦550,000/month suddenly moving ₦3,000,000 is exactly the kind of breach this view is built to surface (+380% in that example), independent of whether any single transaction crossed a hard rule threshold.

10 · Compliance & Governance

Regulatory Reports

STR, SAR, CTR, and AIF filings — with a second signature required by the system, not just by policy.

Tenant AdminCompliance ManagerInvestigator (prepare only)

When a case's decision is REPORT_REQUIRED, a filing gets prepared here — pulling the subject, the suspicious amount, and the grounds directly from the case record, with full evidence lineage back to the alerts and transactions that drove it.

Four-eyes, enforced

Verity will reject an approval if the approver is the same person who prepared the filing — this is checked server-side, not just hidden in the UI, so it can't be bypassed by calling the system directly either. Export packages (JSON or PDF) are only generated once a filing is approved.

11 · Compliance & Governance

Integrations

Your connected source systems and the API keys they push data through.

Tenant Admin

Each source system — core banking, lending, payments, screening — is registered here as an integration with its own scoped API key (shown once, as vrt_..., never again in full). Keys can be rotated or revoked instantly, and each carries explicit scopes (for example events:write) rather than blanket access.

12 · Compliance & Governance

Audit Trail

A statutory, append-only log of every action that mattered.

Tenant AdminCompliance ManagerAuditor

Every decision, status change, and configuration edit is recorded here with the actor, the action, the entity it touched, and a request ID — and it cannot be edited or deleted, by anyone, including a tenant admin. The database role Verity's application runs under is explicitly denied UPDATE and DELETE on this table at the infrastructure level, not just in application logic.

13 · Compliance & Governance

Administration

Your team, your roles, your organization's own settings.

Tenant Admin

Invite teammates by email with a pre-assigned role, review who's active, revoke access instantly when someone leaves, and manage your organization's profile and security settings (including whether MFA enforcement can ever be relaxed — it can't, by design).

Reference

Security & Trust

What's actually enforced underneath the screens above — not just promised.

Dual-layer tenant isolation

Every query is scoped to your tenant twice over — once in application code, once again by PostgreSQL Row-Level Security at the database layer. One institution's data is structurally unreachable from another's, even if an application bug ever tried.

Mandatory MFA, everywhere

TOTP two-factor is required for every account with no opt-out — enforced at login, not left to individual discipline.

Four-eyes on every filing

A statutory report's preparer and approver can never be the same person. Checked server-side, so it holds even against a direct API call.

Append-only audit log

Every decision is permanently recorded with actor, action, and timestamp. The application's own database role is denied the ability to modify or delete audit records, at the infrastructure level.

No eval, no arbitrary code

Detection rules are declarative JSON evaluated by a safe, structured interpreter — never executed as code. A rule author can't accidentally (or deliberately) run anything beyond the condition types Verity defines.

PII masked by default

Emails, phone numbers, and identity fields are masked in every view, for every role, unless explicitly unmasked for a specific, logged reason.

Reference

Glossary

Terms used throughout Verity, in plain language.

Four-eyes segregation

The rule that whoever prepares a filing cannot be the one who approves it — enforced by the system, not just policy.

EDD (Enhanced Due Diligence)

A heightened review decision applied to a customer whose risk profile warrants closer scrutiny than standard KYC.

STR / SAR

Suspicious Transaction / Activity Report — statutory filings submitted to the regulator (e.g. NFIU) when a case's findings require it.

CTR

Currency Transaction Report — a statutory filing for transactions crossing a fixed reporting threshold, regardless of suspicion.

Structuring

Deliberately splitting transfers to stay just under a reporting threshold — one of Verity's built-in rule types detects this pattern directly.

Watchlist signal

A screening hit against a sanctions or watchlist database. Always treated as a signal to investigate — never, on its own, a declaration of guilt.

Row-Level Security (RLS)

A PostgreSQL feature Verity uses as a second, database-enforced layer of tenant isolation, independent of application code.