Customers and suppliers
Prepare sanctions, PEP, wanted-list, identifier, and available research context during onboarding or supplier review.
Solutions / Agentic AML
Give AI Agents Access to Trusted AML Intelligence
Connect compatible AI agents such as Codex and Claude to trusted AML screening through MCP. Automate repetitive compliance workflow steps, reduce screening costs, and recover analyst capacity while your MLRO remains in control.
Business requestScreen this customer, supplier, wallet, vessel, or payment
Agentic orchestrationAuthorised AI agent
01Understand
02Choose tool
03Call
04Organise
Governed connectionMCP + OAuth
Checklynx AML tools
ScreeningSanctions + PEP
IdentifiersWallets + IDs
ResearchMedia + results
Human controlMLRO / ComplianceReview, decide, escalate
MCPA standard connection for compatible AI agents
OAuthAuthorised access in the customer context
EvidenceStructured results with source context
ControlHuman policy, review, escalation, and decisions
Make AML a Capability Your Authorised Agents Can Use
Agentic architecture brings trusted AML checks into the systems where risk enters the business. Instead of moving names and identifiers between screens, teams can ask an authorised agent to call an appropriate Checklynx capability and prepare the result for review.
Reduce workflow friction
Cut navigation, re-entry, result retrieval, and formatting around onboarding, supplier, payment, wallet, and investigation checks.
Use specialist AML intelligence
Ground screening in Checklynx results and source context instead of asking a language model to infer sanctions, PEP, wanted-list, or adverse-media risk.
Keep the MLRO in control
Use agents for coordination and preparation while policy, consequential review, escalation, and final decisions remain with accountable people.
How Checklynx fits into an agentic AML workflow
MCP creates the connection between a compatible AI environment and permitted Checklynx tools. Each layer has a distinct responsibility.
| Layer | Responsibility | Output |
|---|---|---|
| Business workflow | Supplies the customer, company, beneficiary, wallet, vessel, payment, or investigation context. | AML request |
| Authorised AI agent | Understands the request and selects a permitted Checklynx capability. | Tool call |
| Checklynx AML layer | Runs the supported screening or research task and returns structured context. | AML result |
| Compliance review | Applies internal policy, assesses evidence, and decides whether to proceed, investigate, or escalate. | Accountable decision |
Bring governed AML intelligence to the point of risk
One authorised connection can support multiple workflows without requiring every business user to become a screening specialist.
Payments and wallets
Request supported counterparty or identifier checks from payment, beneficiary, payout, or crypto workflows.
Investigations
Retrieve retained screening results and organise source context before substantive analyst or MLRO review.
Operating modelAgent coordinates. Checklynx screens. People decide.
Automation supports the workflow without transferring accountability to the model.
Recover analyst capacity without automating judgement
Agentic AML automation reduces the cost of repetitive screening preparation without replacing compliance professionals. Analysts can spend more time on evidence, ambiguous matches, escalation, and risk decisions.
Less navigation
Bring the AML result into the workflow where the request begins.
Less re-entry
Reuse available names, identifiers, and business context instead of copying them between systems.
Faster review preparation
Organise screening results and source context before the analyst starts substantive review.
Illustrative capacity model3,000 monthly AML tasks × 2 minutes saved = 100 analyst hours
At an illustrative loaded cost of €50 per hour, that represents €5,000 of monthly analyst capacity.
This is an illustration, not a promised Checklynx customer outcome. Actual value depends on workflow volumes, controls, data quality, and measured time saved.What authorised AI agents can do with Checklynx today
The current Checklynx MCP scope is deliberately focused on screening, research, and result retrieval within the authorised customer context.
Screen
Sanctions, PEP, wantedCheck people and companies
Run supported screening for individuals and entities using the permitted Checklynx capability.
Identify
Exact identifiersUse the right specialist check
Check supported identifiers, including wallet identifiers where available, rather than relying on model memory.
Research
Media + retained resultsPrepare the review context
Use adverse-media research and retrieve retained results or recent searches within the authorised customer context.
Governed connectionAccessOAuth + authorisation ToolsNarrowly permitted DecisionHuman controlled
The agent acts within the connected customer's authorised context.
The connection exposes defined AML capabilities rather than unrestricted platform access.
The agent does not autonomously approve KYC, block payments, file reports, or replace policy.
Portal, API, or MCP
Choose the access model based on who initiates the work and how the request is understood.
| Access model | Best for | How it works |
|---|---|---|
| Screening portal | Analyst-led searches, investigations, and review | A person opens Checklynx and performs the work directly. |
| Real-time API | Deterministic onboarding, payment, payout, and backend events | Software makes a predefined call when a known event occurs. |
| MCP | Intent-led work coordinated by an authorised AI agent | The agent interprets the request and selects a permitted Checklynx tool. |
Controlled access for an accountable compliance process
Production use should validate transport, authentication, tool compatibility, tenant scope, evidence retention, and human escalation for the chosen AI environment.
What is Agentic AML?
Agentic AML uses authorised AI agents to coordinate AML work by calling governed screening, research, and retrieval tools while policy, evidence review, escalation, and final decisions remain under organisational control.
Is MCP only for ChatGPT?
No. MCP is an open protocol used by a growing ecosystem of compatible AI agents and applications. Transport, authentication, and tool compatibility should be validated for each environment.
Can an AI agent make the final compliance decision?
The Checklynx MCP surface is designed to provide screening and research capabilities to an authorised agent. The customer's MLRO and compliance team remain responsible for policy, review, escalation, and final decisions.