Using Claude AI with IntelliCompli
IntelliCompli's Model Context Protocol (MCP) integration lets you use Claude Desktop as an AI co-pilot for your AML/CTF compliance workflows. Access 25 powerful tools for customer screening, AUSTRAC reporting, investigations, and risk assessment — all via natural language.
What is the Claude AI integration?
IntelliCompli integrates with Claude Desktop via Model Context Protocol (MCP) — a new standard that lets Claude access external systems securely. Once connected, you can ask Claude natural language questions about your compliance data, and Claude will use IntelliCompli's API to answer them.
Instead of navigating the IntelliCompli dashboard to search customers, review alerts, run screenings, or generate reports, you simply ask Claude: "Show me all critical alerts from this week" or "Screen this name against sanctions lists" or "Generate a TTR for this transaction." Claude handles the API calls, data processing, and result formatting — all while keeping your data secure within your tenant.
Powerful tools covering screening, reporting, investigations, and analytics
Ask Claude questions in plain English — no technical knowledge required
All data stays in your tenant. Claude cannot access other customers' data.
Getting started
Prerequisites
Setup steps
Copy your IntelliCompli MCP URL
Log in to IntelliCompli, go to Settings > Integrations > Claude AI Integration. You'll see your unique MCP server URL. Copy it to your clipboard.
Generate an API key
In the same settings page, click 'Generate API Key' to create a new key. This is used to authenticate Claude with your IntelliCompli account. Keep this secret.
Open Claude Desktop settings
In Claude Desktop, click the menu button in the top-right corner and select 'Settings'. Find the 'Connectors' or 'Integrations' tab.
Add the IntelliCompli connector
Click 'Add New Connector' and paste the MCP URL from Step 1. Enter your API key in the authentication field. Save the configuration.
Test the connection
In Claude, type 'Show me my IntelliCompli customers'. If Claude responds with a list of customers, the connection is working.
Connection is secure
The MCP protocol uses encrypted connections. Your API key is sent securely and is never stored in Claude. You can revoke the key at any time from IntelliCompli Settings.
Available tools (25 total)
All IntelliCompli tools are available through Claude. Tools are organized by function. Use natural language to access any of them — Claude will call the right tool based on your request.
Customer Management
list_customers
Search and filter customers by email, name, risk level, or screening status
get_customer
Retrieve full customer profile including KYC data, risk rating, and history
create_customer
Create new customer record with initial KYC and risk assessment
Sanctions & PEP Screening
screen_sanctions
Run real-time sanctions screening against DFAT and UN databases
screen_pep
Check against Politically Exposed Person (PEP) databases
Alerts & Investigations
list_alerts
List compliance alerts filtered by severity (critical, high, medium) and status
get_alert
Get detailed alert information including triggering rule and evidence
resolve_alert
Resolve or dismiss alerts with documented reasoning
list_cases
List investigation cases with filtering by status and assigned team
get_case
Get investigation case details, timeline, and findings
create_case
Open new investigation case for complex compliance matters
update_case
Update case status, add findings, and manage case lifecycle
Transactions
list_transactions
List transactions with filtering by amount, date, risk level, or TTR status
get_transaction
Get transaction details including customer, amount, and risk assessment
create_transaction
Record new transaction for risk assessment and reporting
Risk Assessment
assess_customer_risk
Trigger automated risk reassessment across 11+ risk factors including customer profile, transaction history, and sanctions screening
AUSTRAC Reporting
generate_ttr
Generate Threshold Transaction Report for transactions >= $10,000 AUD
generate_smr
Generate Suspicious Matter Report with evidence and investigation details
generate_ifti
Generate International Funds Transfer Instruction for cross-border transfers
list_smr_reports
View all submitted Suspicious Matter Reports with status and timestamps
Analytics & Compliance
get_analytics
Get overview metrics: customers screened, alerts triggered, reports filed
list_audit_logs
Full audit trail of all platform actions with user attribution and timestamps
list_edd_investigations
List all Enhanced Due Diligence investigations with status
get_edd_investigation
Get details of a specific Enhanced Due Diligence investigation
list_beneficial_owners
List Ultimately Beneficial Owners (UBOs) for business customers
Example prompts
Here are real-world examples of what compliance officers ask Claude when working with IntelliCompli. Use these as inspiration for your own queries.
Daily Operations
Show me all critical and high severity alerts from this week
List customers with high risk level
What transactions were flagged for review today?
Display all alerts with pending resolution
Investigations
Get the full profile for customer cus_abc123 including their screening history
Create a new investigation case for suspected structuring by customer cus_xyz
Show me all transactions for customer cus_abc123 over $5,000
What's the timeline of alerts for this customer?
Screening
Screen John Smith born 1985-03-15 against sanctions lists
Run a PEP check on Sarah Johnson
Check if this business name appears in any sanctions databases
Reporting
Generate a TTR report for this month
Show me all SMRs filed this quarter
What's the analytics overview for the last 30 days?
List all pending AUSTRAC reports
Audit & Compliance
Show the audit trail for customer cus_abc123
List all open EDD investigations
Who are the beneficial owners of business customer cus_biz456?
What's the complete record of actions taken on this case?
Best practices
To get the most out of Claude + IntelliCompli, follow these practices:
Start with read operations
Begin by querying data (list_customers, get_alert) before making changes. This helps you understand the situation before taking action.
Review before submitting reports
Claude will show you draft reports before submission. Always review AUSTRAC-bound reports (TTR, SMR, IFTI) before confirming.
Use specific IDs for investigations
Include customer IDs (cus_*), case IDs, or transaction IDs when investigating specific matters. This ensures accuracy and prevents mistakes.
Combine multiple queries
Layer queries for deeper analysis: get customer profile, list their transactions, check screening results, then assess risk.
Confirm before write operations
Claude will request confirmation on any action that modifies data (create_case, update_case, resolve_alert). Review the action details before confirming.
Use analytics for dashboarding
Use get_analytics to generate real-time dashboards of compliance metrics for weekly reporting or management reviews.
Security, permissions & audit
Your API key is secure
Your API key is used only to authenticate Claude's requests to IntelliCompli. It is never logged, stored in Claude's memory, or shared with Anthropic. If you suspect the key has been compromised, immediately revoke it in Settings and generate a new one.
Claude respects your permissions
Claude cannot bypass your role-based permissions in IntelliCompli. If you don't have permission to view a customer or file a report, Claude cannot access it either. All API calls are authenticated using your API key and subject to your account's permission boundaries.
All actions are logged
Every action Claude takes on your behalf is logged in IntelliCompli's audit trail. The log shows what action was taken, by which user (the Claude client), at what time, and what data was accessed. You can review the full audit history at any time.
Data stays in your tenant
IntelliCompli is a multi-tenant platform. Claude can only access data belonging to your tenant. Cross-tenant data access is impossible — customers, alerts, reports, and configurations are completely isolated from other organizations.
Frequently asked questions
Does Claude see the AI conversation?
No. Claude has no memory of previous conversations. Each conversation starts fresh. Your prompts and Claude's responses are not stored in IntelliCompli.
Can Claude create customers or modify existing ones?
Yes, Claude can create new customer records via the create_customer tool. However, Claude will always ask for confirmation before making changes. You must approve any action.
What if I accidentally ask Claude to do something wrong?
Claude has safeguards built in. It won't bypass permissions, overwrite critical data, or take actions that contradict compliance rules. If Claude detects a potentially problematic request, it will ask for clarification.
Can I use Claude to generate AUSTRAC reports?
Yes. Claude can generate TTRs, SMRs, and IFTIs using the generate_ttr, generate_smr, and generate_ifti tools. Claude will show you a draft for review before submission.
How do I disconnect Claude from IntelliCompli?
Go to Settings > Integrations > Claude AI Integration and click 'Revoke API Key'. This immediately disconnects Claude and invalidates the key. You can generate a new key anytime.
Is there a cost for using Claude with IntelliCompli?
Claude API usage is subject to Anthropic's pricing. The IntelliCompli integration itself is included in your IntelliCompli plan at no additional cost. Contact sales for details.
Disclaimer: This content is provided for general informational purposes only and does not constitute legal, financial, or professional compliance advice. While we endeavour to keep this information accurate and up to date, legislation and regulatory guidance change frequently. You should seek independent legal or compliance advice specific to your circumstances before acting on any information in this guide. See our Privacy Policy for how we handle your data.
Start using Claude AI with IntelliCompli
Connect your IntelliCompli account to Claude Desktop in minutes. Use natural language to automate screening, reporting, investigations, and risk assessment.