RiskPulseRequest a Demo

The platform

One governed platform for financial crime compliance.

RiskPulse plans the file, gathers and validates the evidence, runs the client interaction, and delivers a result an examiner can read. Every step is configured by your policy and adjudicated by your people.

Request a DemoSee how it configures
Abstract composition of a governed process
PolicyEvidenceSign-off

The compliance program

Revenue at one end.
Wasted effort
at the other.

Onboarding sits at the revenue end, while investigation sits at the cost end, of the financial crime compliance value chain — between them is the monitoring system you have already spent years and budget tuning. We leave that alone, accelerate one end, and compress the other.

The compliance program: onboarding, monitoring and investigation, wrapped by governance and quality assuranceGOVERNANCEAudit trail, explainability, complete case recordsOnboardingCommercial and capital marketsOwnership, screening, EDDPeriodic and event-driven reviewTransaction monitoringYour existing systemRules and modelsRiskPulse is agnosticInvestigationAlert disposition and case workRegulatory reportingCorporate fraud mattersQUALITY ASSURANCESampling methodology, independent agentic case assessment

Input

We do not replace
your monitoring.

Most institutions have spent years tuning a transaction monitoring system. RiskPulse is deliberately agnostic to it — the agents take the alert it raises and investigate from there, returning the dispositioned case to the system your team already works in.

Work arrives from where it already lives.

  • Monitoring alerts — from your existing transaction monitoring system, whatever it is
  • Onboarding requests — new relationships (CRM), periodic reviews, and event-driven triggers
  • Referrals — audit findings, whistleblower reports, and management escalations
  • Connected sources — screening and registry data, adverse media, mail, document stores, chat and drives
  • Your case management — ours or yours; outcomes are written back to where your team works
RiskPulse data ingestion view

How it runs

Plan. Examine. Report.

Three coordinated agents, and each answers for its own work. A reviewer can interrogate the plan, the evidence, and the conclusion independently — accept, amend, or reject each without disturbing the rest.

Plan, Examine and Report agents, each followed by a configurable review gateAGENT 01PlanReads the case, applies the templateBuilds the full task listAGENT 02ExamineGathers and validates evidenceCites every sourceAGENT 03ReportScores the risk dimensionsDrafts the narrativeReview gateAccept, amend or rejectReview gateAccept, amend or rejectSign-offApprove and closeEach gate is a policy setting — every action, override and approval is written to the audit record

Configuration

Your policy
is the program.

Every case carries a domain and a category. Together they select a policy — and that policy configures what the agents do, what they may touch, how the risk is scored, and what the finished report looks like. Nothing runs on a generic definition of good practice.

Domain and category select a case policy, which configures the plan template, risk dimensions, tool access and report templateDOMAINCorrespondent banking, retail, capital marketsCATEGORYFraud, name screening, transaction investigationYour policies govern the platformOne policy per domain and category pairPlan template — the tasks the agent may generateRisk dimensions — scores, weights, override rulesTool access — which systems the agent may reachReport template — sections, format, house style
Plan template
Each use case has its own plan template. The Plan Agent works from it, tuning the task list to the facts of the specific case rather than generating a plan from scratch.
Risk dimensions
The dimensions scored, their weights, the composite bands, and the override rules that force an outcome regardless of score — all defined by you, per case type.
Tool access
Which systems each agent may reach is a policy setting, not an installation default. An agent working a screening alert has the screening and registry sources that case requires and nothing else — scoped per case type, enforced at the platform, and recorded in the audit trail alongside every other action.
Report template
You define the sections, the order, the house style, and the fields. The Report Agent writes into your template, not a generic summary format.
Review gates
Where a person must intervene is configurable. A low-risk file can run through to sign-off; a higher-risk case type can require review at every stage.

The same engine, different policy

One structure.
Different questions.

This is why a single platform can run a name screening disposition and an institutional onboarding file. The scoring structure never changes — weighted dimensions, a composite band, override rules. What changes is which dimensions are being scored, and that is yours to define.

Three case types share one scoring structure but use entirely different risk dimensionsALERT INVESTIGATIONParty riskTransaction riskJurisdiction riskBusiness contextAlert riskMERCHANT FRAUDMerchant riskTransaction riskControl riskCollusion riskDocumentation riskINSTITUTIONAL ONBOARDINGSanctions and PEPBeneficial ownershipManager regulatorySource of capitalOperational conductSame structure throughout — weighted score, composite band, override rules. The dimensions themselves are policy.

Output

A defensible record,
not a summary.

Findings, actions and outputs are assembled into a complete case record. Every conclusion carries its reasoning chain and a citation to the source document, so the file can be re-read, questioned, and defended long after the analyst who closed it has moved on.

  • Scored determination — each dimension rated and weighted, with the composite band and the rule that drove the outcome
  • Written rationale — why the conclusion was reached, in your template and house style
  • Supporting evidence — every finding tied to the document, record or screening result behind it
  • Assessment limits — what could not be established, and why, stated rather than glossed over
  • Complete audit trail — agent actions, human overrides and approvals, in sequence
RiskPulse assembled case record screenshot

Governance and quality

Built to be
defended.

Two separate controls sit around the work: governance, which makes each case explainable after the fact, and quality assurance, which tests whether the work was any good in the first place.

Governance

Explainable per case

Every case carries its full reasoning chain and source citations. What the agent did, what a person changed, and who approved it — recorded in sequence and reconstructable on demand.

Quality control

Sampling methodology

A sampling methodology is built into every case type, so your quality team can draw and assess a representative sample rather than reviewing at random or reviewing everything.

Quality assurance

Independent case assessment

A separate set of agents assesses case quality end to end against your methodology — a second opinion on the work, not a second run of it.

On AI-generated content

Every RiskPulse output states plainly that it was produced with the support of large language models, that models can err, and that findings must be validated by a qualified professional before disposition. We would rather volunteer that than be asked for it.

Technology and architecture

Your tenancy.
Your keys.

RiskPulse is a containerised, cloud-native platform that deploys into an isolated tenancy. It is designed so that the security review is short.

  • Tenant isolation — a dedicated environment per institution, with no shared processing of your data
  • No public ingress required — access over private connectivity from your own network
  • Your encryption keys — data at rest encrypted under keys you create, own and can revoke
  • Role-based access control — investigator, reviewer and administrator roles, with segregation of duties enforced
  • Controlled egress — outbound access restricted to the external sources you approve
  • Portable — the environment can transition into your own cloud ownership and governance
Users reach an isolated tenancy over private connectivity, with controlled outbound access to approved external sourcesYour networkInvestigatorsReviewers, adminsISOLATED TENANCYEdgeNo public ingress requiredApplicationRole-based access controlData and modelsEncrypted under your keysApproved externalsourcesControlled egress only

Solutions

Six configurations
of one platform.

Each is the same engine under a different policy. Ninety percent of our focus is financial crime in financial services; corporate fraud is covered where the work is investigative in the same way.

Onboarding

Onboarding

Commercial and capital markets files, from request through approval — documents, ownership, screening, and periodic review.

Explore →Investigations

Correspondent Banking

Respondent relationships, nested activity, and questionnaire review, evidenced against the correspondent standard.

Explore →Investigations

Trade-Based Money Laundering

Trade documents read across languages, pricing tested against benchmarks, routes and counterparties screened.

Explore →Investigations

AML Investigations

Monitoring alerts dispositioned with a written rationale, cases investigated, narratives drafted for reporting.

Explore →Investigations

Corporate Fraud

Misappropriation, internal matters and whistleblower referrals — scoped, evidenced, and written up.

Explore →Screening

Name Screening & Sanctions

Name, sanctions, and adverse media screening handled from alert to disposition, with evidence attached and rationale written to your policy standard.

Explore →

See it on your own file

Bring a case.
We will run it.

Pick a real file or a closed case and see how RiskPulse plans it, evidences it, and writes it up under your policy.

Request a Demo