CrimeTracer™
AI Chatbot – Agency FAQ

The CrimeTracer AI Chatbot is a natural language interface that allows users to query their agency’s data using plain English instead of traditional dropdowns or forms. It is designed to help law enforcement professionals quickly access relevant information using simple questions or commands.

The chatbot is powered by Claude 3.1, one of the most advanced and secure large language models available. It has been configured specifically for law enforcement use cases within the CJIS-compliant CrimeTracer platform.

• Converts natural language queries (like “Show me all incidents involving a red pickup truck and a white male aged 30 to 35 with brown eyes.”) into SQL queries that are executed against your agency’s CrimeTracer database.
• Makes it faster and easier to search your data without needing to navigate complex forms.
• Helps users locate incidents, entities, patterns, and context efficiently through conversational search.

• It does not modify, delete, or alter any data.
• It does not generate assumptions or conclusions about the data. All responses are based on the factual contents of your database.
• It does not send data outside the CJIS boundary or log conversations outside your secure environment.

Yes. The chatbot has been implemented with strict adherence to CJIS compliance requirements. The model processes queries within a secured environment, and no agency data is sent to external servers or used to train the model.

The chatbot is available to licensed users of CrimeTracer. Role-based access controls ensure users can only query data they are already authorized to see within the platform.

To activate the CrimeTracer AI Chatbot:

Send a request to CrimeTracer Support at crimetracer.support@soundthinking.com. A support specialist will walk you through activation and permissions.

Yes. CrimeTracer support can enable the chatbot for a limited pilot group within your agency to evaluate the feature before making it available more broadly.

Yes. All AI-generated SQL queries and responses are logged in accordance with agency auditing policies and CJIS security standards.