KYC/AML Compliance Agent · Use case
How can a bank use an AI agent to prepare KYC and AML alerts for analysts on its own infrastructure?
A KYC and AML compliance agent takes a screening or monitoring alert, gathers the customer record, the transactions and the screening result, and writes up the evidence for an analyst. The analyst decides whether to close or escalate the alert: the agent closes nothing and files no report. It runs on your own infrastructure, so customer and transaction data stay inside the bank.
Alert after alert.
Most of the time is gathering.
- ✗ 01Analysts spend most of an alert gathering the same records from several systems before they can judge it.
- ✗ 02Alerts hold customer and transaction data that must not leave the bank, and every decision must be explained to a regulator.
Where it stops for a person
Always, before an alert is closed or escalated.Always. An analyst closes or escalates every alert, and a person decides whether to report. The steps the agent took are logged.
It gathers the evidence.
An analyst decides.
Step · 01
An alert arrives from your screening or monitoring system.
Step · 02
The agent reads the customer record and the transactions behind the alert.
Step · 03
It collects the screening result and any earlier alerts on the same customer.
Step · 04
It writes a summary that cites each record it used.
Step · 05
An analyst reviews the summary and closes or escalates the alert.
Reads from · connectors
Database
Your own APIs
SharePoint
The agent connects to a Database for customer and transaction records, and to your screening and monitoring systems through your own APIs. Your procedures are read from SharePoint.
Acts through · MCP servers
Nothing. It writes up the evidence for the analyst.
What you set up
- Connect the customer and transaction records.
- Connect your screening and monitoring systems.
- Write test alerts with the outcome your analysts reached.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does the agent close alerts?
No. It assembles the evidence. An analyst closes or escalates.
Does it replace our screening system?
No. It reads the results of the system you already use.
Can we show a regulator what it did?
Yes. Each summary cites its records, and the steps are logged.