Actions and Navigation

The AI Assistant does not just answer questions -- it helps you act on the answers. Many responses include interactive action buttons that let you navigate to relevant pages, trigger remediations, run scans, and more, all without leaving the chat panel.

Action Buttons

Action buttons appear at the bottom of assistant responses when there is a clear next step you can take. They are styled as clickable buttons within the chat message, making it easy to go from insight to action in a single click.

How Action Buttons Work

1. The assistant analyses your question and its response.
2. If there is a relevant action you can take, one or more action buttons are included in the response.
3. Click a button to trigger the action immediately.
4. The result of the action depends on its type -- you may be navigated to a page, prompted to confirm a remediation, or see a scan begin.

info

Action buttons are contextual. They only appear when relevant to the conversation. Not every response includes action buttons -- informational answers that do not have a clear next step will simply provide the information without buttons.

Action Types

The AI Assistant supports the following action types:

Navigate

Navigate buttons take you directly to the relevant page in the Guardian Pro console, often with filters pre-applied.

Examples:

Button Label	Where It Goes
"View in Action Centre"	Action Centre filtered to the specific finding discussed
"Open Cost Dashboard"	Cost Intelligence page
"View Compliance Dashboard"	Compliance Dashboard filtered to the relevant framework
"View Architecture Map"	Architecture Advisor with the relevant resource highlighted
"See All Findings"	Action Centre with appropriate severity or service filters

Navigate buttons are the most common action type. They bridge the gap between getting an answer in the chat and seeing the full detail on the appropriate page.

tip

When the assistant mentions a specific finding, risk, or resource, look for a navigate button to jump directly to its detail view. This is faster than manually navigating and applying filters.

Remediate

Remediate buttons let you trigger a fix for a specific finding directly from the chat conversation.

How remediation from chat works:

1. You ask about a finding or ask the assistant to fix something.
2. The assistant explains what the remediation will do and includes a "Fix Now" or "Remediate" button.
3. Clicking the button opens the remediation preview modal, showing exactly what will change.
4. You review the preview and confirm to proceed.
5. The remediation executes, and you can monitor its progress.

caution

Remediation actions always require your explicit confirmation. Clicking a remediate button opens a preview -- it does not immediately apply changes. You always have the opportunity to review what will happen before confirming.

Example conversation:

You: "I have an S3 bucket with public access. Can you fix it?"

Assistant: "Yes, the finding s3_public_access affects bucket my-app-uploads. Guardian Pro can block public access by enabling the S3 Block Public Access settings. This is a reversible change."

[Fix Now] [View Finding]

Run Scan

Scan buttons let you trigger a new infrastructure discovery and security scan.

When they appear:

• When you ask about the current state of your environment and the most recent scan is older than expected.
• When the assistant suggests running a new scan to get updated data.
• When you explicitly ask to scan your environment.

Clicking "Run Scan" starts the scan immediately. You can monitor progress from the Dashboard or continue chatting while the scan runs in the background.

View Health

Health buttons navigate to the dashboard health score view, often appearing when the assistant discusses your overall infrastructure health or contributing factors.

View Risks

Risk buttons navigate to the Architecture Advisor's risk radar, filtered to the specific risk type or severity discussed in the conversation.

Simulate

Simulation buttons launch a failure simulation for a specific resource discussed in the conversation. This takes you to the Architecture Advisor's failure simulator with the resource pre-selected.

Example conversation:

You: "What happens if my production database fails?"

Assistant: "A failure of RDS instance prod-db would affect 12 dependent resources including your API containers and cache clusters. The estimated impact covers 45% of your infrastructure."

[Run Full Simulation] [View Architecture Map]

Generate Infrastructure

Infrastructure buttons navigate to the Infrastructure Wizard, optionally preloaded with context from the conversation.

Example conversation:

You: "I need to set up a new staging environment with the same architecture as production"

Assistant: "I can see your production environment uses containers with a PostgreSQL database, Redis cache, and S3 storage. You can use the Infrastructure Wizard to generate a staging-optimised template based on this architecture."

[Open Wizard]

Open Wizard

Similar to generate infrastructure, this button opens the Infrastructure Wizard start page for creating new infrastructure from scratch.

Follow-Up Suggestions

In addition to action buttons, the assistant often includes follow-up question suggestions at the end of its response. These appear as clickable text links that pre-fill your next message.

Example:

Assistant: "Your top 3 critical findings are related to IAM, S3, and RDS..."

Follow-up suggestions:

• "Tell me more about the IAM finding"
• "Can these be automatically remediated?"
• "What is the compliance impact of these findings?"

Clicking a follow-up suggestion sends that question immediately, making it easy to continue the conversation without typing.

Action Security

All actions initiated through the AI Assistant follow the same security model as actions taken directly in the Guardian Pro console:

• Permission checks -- Actions are validated against your user permissions. If you do not have remediation permissions, remediate buttons will not appear.
• Confirmation required -- Destructive or modifying actions always require explicit confirmation through a modal dialog.
• Audit trail -- All actions triggered from the assistant are logged in the same audit trail as actions taken from the console directly.
• Account scope -- Actions apply only to the currently selected AWS account.

info

The assistant is aware of your permission level. If you ask about remediation but do not have the security:remediate permission, the assistant provides the remediation steps for manual execution instead of offering an automated fix button.

Actions and Multi-Account

Action buttons always operate in the context of the currently selected AWS account. If you switch accounts:

• Action buttons from previous responses are no longer clickable (they show the account context they were generated for).
• New responses generate action buttons for the current account.

This prevents accidentally applying an action to the wrong account.

Combining Chat and Console

The AI Assistant is designed to work alongside the rest of the Guardian Pro console, not replace it. A common workflow is:

1. Ask a broad question in the chat to understand the current situation.
2. Click an action button to navigate to the relevant page.
3. Use the full console interface to review details, apply filters, and take action.
4. Return to the chat for follow-up questions or to explore a different topic.

The chat panel stays open as you navigate between pages, so you can reference the assistant's answers while working in the console.

Next Steps

• Starting Conversations -- Learn how to frame questions for the best results.
• What to Ask -- Browse example questions across all categories.
• Conversation History -- Manage your conversations and revisit previous interactions.