Using the Chatbot in Investigative Workflows
AI Chat can support investigative workflows by helping you summarize evidence, identify connections, and surface relevant details across large datasets. Use it as an exploratory tool to guide analysis, but always verify important findings directly in the source material.
Start with an overview
A common first step is to ask AI Chat for a high-level summary of the evidence you are reviewing.
Example scenario You are reviewing WhatsApp documents containing private luxury vehicle sales invoices linked to a suspect’s phone number.
- Open AI Chat.
- Add the relevant folder or documents as context.
- Ask a focused question such as:
“Summarize all references to private vehicle sales in these WhatsApp documents and identify the main individuals involved. Provide citations.”
Use the response to:
- identify recurring names, companies, or identifiers (phone numbers, IBANs, license plates)
- extract transaction details such as amounts, dates, and locations
- detect patterns in conversations (repeated sellers, brokers, intermediaries)
- flag possible financial relationships that should be modeled as records
Validate the response
Always verify AI-generated insights before relying on them.
- Review the references section in the response.
- Open each cited file or record.
- Confirm names, dates, and amounts directly in the source evidence.
Deepen the analysis with Records and Link Charts
After the initial overview, move from summarization to structured investigation.
Typical steps include:
- creating or opening relevant Records (people, organizations, vehicles)
- adding or reviewing Relationships supported by evidence
- opening a Link Chart to visualize connections between entities
You can use follow-up questions in AI Chat to narrow the scope or test investigative hypotheses. Example prompts:
- “Which relationships connect these records between 2023-01-01 and 2023-06-30? Provide citations.”
- “List the phone numbers and IBANs appearing in both invoices and chat messages, grouped by source.”
- “Which entities appear in multiple files, and what evidence supports each connection?”
A practical workflow loop is:
- Ask a focused question.
- Review citations.
- Update records or the Link Chart.
- Repeat.
AI Chat vs. AI Find by Question
Use AI Chat and AI Find by Question for different purposes during an investigation.
| Use AI Chat for | Use AI Find by Question for |
|---|---|
| Quick overview of evidence | High-precision extraction |
| Cross-document reasoning | Yes/No checks per document |
| Exploratory questions and hypothesis building | Structured outputs (columns per question) |
| Identifying potential connections | Document-by-document validation |
| Suggesting investigative leads | Exportable tables (CSV) |
Example workflow
- Use AI Chat to identify key patterns and determine which fields you need to extract.
- Switch to AI Find by Question and define questions as columns, for example:
- “Is a vehicle sale mentioned?”
- “What is the sale amount (EUR)?”
- “Who is the seller?”
- “Who is the buyer?”
- “Which phone numbers or IBANs appear?”
- Run the extraction and review the resulting table.
- Export the results and attach them to the workspace as investigation evidence. For advanced usage, see the AI Find by Question section of this guide.
Verification checklist
Whether you use AI Chat or AI Find by Question, always verify the results before acting on them.
- Review the references section.
- Ask for citations when responses contain key claims.
- Open the original files or records to confirm details.
- Treat conclusions as unverified until the source evidence confirms them.