Accessing Protegrity Agent UI
Accessing the Dashboard
Access the PPC using the FQDN provided during the installation process.
Enter the username and password for the admin user to log in.
If Protegrity Agent is installed, then the Protegrity Agent dashboard appears. If Protegrity Agent is not installed, then the Insight Dashboard is available.
Afetr logging in successfully, a Welcome message is displayed prompting users to start a conversation.
The left panel contains Ongoing Conversation and Chat History icons.
Click Chat History to refer to previous requests and results.
The Protegrity Agent version is displayed at the bottom-right of the page. Currently, Protegrity Agent is on version 1.0.
Starting a Conversation
Get started with Protegrity Agent by typing a question in the Start your conversation here… textbox. The Protegrity Agent breaks down the request into actionable steps and executes them, providing live updates in the Canvas.
Once the Agent starts processing the query, the stop icon is enabled. Use this icon to stop the Agent from proceeding with the request.
Interacting with the Agent
The following are the best practices while interacting with the agent.
Be Specific
Provide clear instructions. Include resource names, values, and desired outcomes.
| Less effective | More effective |
|---|---|
| “Set up some security.” | “Create a policy named PCI_Compliance with a tokenization rule for the CC_Token data element.” |
| “Fix the policy.” | “Update the PCI_Compliance policy to change the masking rule for SSN_Token from full mask to partial mask.” |
Understanding the Canvas
The attributes displayed in the canvas response along with their description are listed below.
| Attribute | Description |
|---|---|
type | Items of an agent response can be one of several types: - TextResponse: The agent’s text output - AgentPlan: A plan created by the agent - ExecutedStep: An individual step in the plan - FeedbackRequest: A request for user feedback - Attestation: Evidence or reasoning from the agent |
id | Unique identifier for this conversation |
claim | The claim or assertion this evidence supports |
confidence_level | Confidence level (0-1) in the validity of this evidence |
evidence_data | The actual evidence data - source: Source of the evidence - collected_at: When the evidence was collected - data: The evidence payload |
evidence_type | Type of evidence being provided |
verification_method | Method used to verify or collect this evidence |
created_at | When the conversation was created |
If the response from the Agent is text information only, the Canvas does not display any output.
Feedback
Was this page helpful?