Skip to main content

Tips & Best Practices

Get the most out of the Tripl-i AI Chatbot with these practical tips.

Writing Effective Queries​

Be Specific When You Can​

The more specific your question, the more precise the answer.

Instead of...Try...
"Show me servers""Show me Windows servers in production"
"Any vulnerabilities?""What critical vulnerabilities affect our SQL servers?"
"What about compliance?""What's our SOX compliance status for the finance department?"

Use Natural Language​

You don't need special syntax or keywords. Ask questions the way you would ask a colleague.

  • "Which servers haven't been scanned recently?"
  • "How many workstations are running Windows 10?"
  • "What depends on the DNS server?"

Use Follow-Up Questions​

The chatbot remembers your conversation. You don't need to repeat context.

  • First: "Show me all database servers"
  • Then: "Which ones have critical vulnerabilities?" (the chatbot knows you mean database servers)
  • Then: "How about the ones in London?" (narrows further)

Reference CIs by Name​

Use the exact CI name or hostname when asking about specific systems:

  • "Tell me about PROD-SQL-01"
  • "What's installed on WEB-SERVER-03?"
  • "Show dependencies for MAIL-PRIMARY"

Getting the Best Results​

Start Broad, Then Narrow​

Begin with a general question to understand the landscape, then drill into specifics:

  1. "How many servers do we have?" (get the big picture)
  2. "How many are running Linux?" (narrow by OS)
  3. "Which Linux servers are in production?" (narrow by environment)
  4. "Do any of those have critical vulnerabilities?" (check security)

Use the Page Context​

When you're viewing a specific CI in the CMDB, the chatbot knows which asset you're looking at. Use this:

  • "What's connected to this server?" (no need to name it)
  • "Any recent events?" (checks the current CI)
  • "Show me installed software" (for the CI you're viewing)

Ask for Summaries​

The chatbot can summarize complex information:

  • "Give me a summary of our infrastructure health"
  • "Summarize the compliance status across all frameworks"
  • "What are the top risks in our environment?"

Ask for Comparisons​

Compare different parts of your infrastructure:

  • "Compare the London and Frankfurt datacenters by server count"
  • "Which department has the most compliance issues?"
  • "What's the difference between our production and staging environments?"

Understanding Responses​

Data Freshness​

Chatbot responses come from your live CMDB data. The accuracy depends on:

  • Discovery scan frequency — More frequent scans mean more current data
  • Last scan timestamp — Check when a CI was last scanned if data seems outdated
  • Manual updates — Changes made through the UI are reflected immediately

Tenant Isolation​

The chatbot automatically filters all queries to your tenant. You only see data you have access to. There is no way to query another tenant's data through the chatbot.

Confidence and Limitations​

The chatbot is excellent at:

  • Querying structured data (servers, software, relationships)
  • Counting and filtering CIs
  • Looking up specific CI details
  • Compliance status queries
  • Vulnerability lookups
  • Dependency mapping

It may be less precise with:

  • Predictive questions ("Will this server fail next week?")
  • Questions about data that hasn't been discovered yet
  • Very complex multi-step analyses requiring human judgment

If you get an unexpected answer, try rephrasing your question or being more specific about what you're looking for.

Write Operations​

When you ask the chatbot to make changes (update a CI field, change a status, add a tag), it will:

  1. Show you exactly what it plans to change
  2. Wait for your explicit confirmation
  3. Only proceed after you approve

This prevents accidental modifications. Review the proposed changes carefully before confirming.

Keyboard Shortcuts​

ShortcutAction
EnterSend your message
Shift + EnterAdd a new line without sending
EscClose the chatbot panel

Common Questions​

Q: Can the chatbot access all my data?

A: The chatbot can query anything in your CMDB that your user account has access to. It respects the same tenant isolation and permission boundaries as the rest of the platform.

Q: Are my conversations stored?

A: Conversation history is maintained for your current session to enable follow-up questions. Conversations are not permanently stored or used to train AI models.

Q: Can I use the chatbot in languages other than English?

A: The AI model understands multiple languages. You can ask questions in your preferred language and receive responses accordingly. Infrastructure data (CI names, OS versions) will appear in their original format.

Q: What if the chatbot gives an incorrect answer?

A: If a response seems wrong, try rephrasing your question with more specific details. The accuracy depends on the completeness of your CMDB data — if a CI hasn't been scanned or updated, the chatbot may not have the latest information.