Chat Assistant (AI)

The Chat Assistant answers everyday questions about your cloud estate in plain language. Ask about your costs, resources, optimization recommendations, and resource pools, and it replies with the specific figures from your own data. There is no query syntax to learn. You type a question the way you would ask a colleague, and the assistant finds the answer and presents it back as a short, data-driven summary.

Opening the assistant

You can open the Chat Assistant in two ways:

  • Chat Assistant in the main navigation menu, towards the bottom of the menu.

  • The chat bubble in the bottom-right corner of many screens, which opens the assistant in a panel without taking you away from the page you are on.

Managing conversations

Your questions are grouped into conversations, so you can keep separate lines of inquiry apart and come back to them later.

  • New Chat starts a fresh conversation. Use it when you move on to an unrelated topic.

  • A conversation contains one or more messages, building up as you ask follow-up questions.

  • Favorite a conversation to keep it close to hand for quick access.

  • Delete a conversation you no longer need.

What you can ask

The assistant understands questions across several areas of your cloud data. The examples below show the kinds of question it handles. You are not limited to these exact phrases, so ask in your own words.

Costs

  • "What are my top cost drivers?"

  • "How much did we spend on AWS last month?"

  • "Show me cost trends for the past quarter."

  • "Break down costs by region" or "by service."

  • "Give me a daily breakdown of spend this month."

Resources

  • "Show me unused resources."

  • "What resources are running in EU-West?"

  • "List all virtual machines by cost."

  • "How many stopped instances do we have?"

  • "Tell me about resource abc-123."

Recommendations and savings

  • "What cost optimization recommendations do we have?"

  • "Which resources should we right-size?"

  • "Show me our potential savings."

  • "What are our high-priority recommendations?"

Resource pools

  • "What resource pools are configured?"

  • "Show me pool utilization."

  • "Which pools are over budget?"

Governance and scheduling

  • "Show me any constraint violations."

  • "What power schedules are in place?"

The assistant also handles follow-up questions within a conversation, so you can refine an answer without repeating yourself. It can combine steps too, for example "Show me unused resources and how much we would save by removing them."

Creating virtual tags

As well as answering questions, the assistant can create a virtual tag for you. Describe the tag you want in plain language and the assistant builds the tag definition, saving you a trip to the tagging screens.

Creating a virtual tag is the only change the assistant can make to your data. It can create a new virtual tag, but it cannot modify or delete existing tags or resources. Everything else the assistant does is read-only.

Tracking usage

The Usage view shows how the Chat Assistant is being used. It reports the conversations and messages, together with the credit usage consumed, broken down per user or per conversation. Use it to see who is getting value from the assistant and to keep an eye on consumption.

Tips for better answers

  • Be specific. "Show me AWS EC2 costs for March" works better than "show me some costs."

  • Include a time period for any cost question.

  • Name the provider when you want to filter to a single cloud.

  • Ask follow-up questions. The assistant remembers the context of the current conversation.

  • Ask for a format. Request a table, a summary, or a comparison if you want the answer shaped a certain way.

Extending the assistant with your other products

You can connect the Chat Assistant to your other Certero products so it can answer questions that span more than one of them, giving you cross-product answers from a single conversation. This is set up by an administrator on the Configuration > AI MCP Integration page. See that page for details.