Ask Zia: The super agent for IT analytics

Ask Zia is the agentic analytics assistant in Analytics Plus Cloud. It interprets the outcome you describe, coordinates the specialized agent best suited to each part of the request, and works through the analytical steps needed to produce a result. IT teams can use Ask Zia to prepare data, create and refine reports and dashboards, investigate changes, and identify possible next steps while retaining control over the output.

You can describe a goal in natural language instead of limiting the interaction to a single question and answer. For example, ask Zia to investigate the week's incident surge, compare application availability across business services, build and refine a service desk dashboard, or determine the factors behind a change in mean time to resolution.

Specialized agents in Ask Zia

Ask Zia coordinates specialized agents across the analytical workflow. It interprets the request and directs it to the capability best suited to the task.

  • Analysis Agent: Examines IT data for trends, patterns, anomalies, contributing factors, and possible next steps.
  • Visualization Agent: Creates and updates reports, dashboards, charts, and summary widgets from natural-language instructions.
  • SQL Agent: Generates, explains, optimizes, and troubleshoots SQL queries for Query Tables.
  • Formula Agent: Creates formula columns and aggregate formulas from natural-language requirements.
  • Data Preparation Agent: Helps build data pipelines and prepare data for analysis.
  • DSML Coding Agent: Assists with Python code in Code Studio for data transformation, analysis, and machine learning.
  • Slideshow Agent: Builds and refines narrative slideshows from reports and dashboards.

Access and permissions

Administrators, shared users, and users assigned custom roles can work with Ask Zia subject to the data permissions defined for their role. Administrators can manage the AI services and capabilities available to users from the GenAI Configuration page.

Chat history

Ask Zia automatically saves conversations so you can return to earlier questions and responses. Select a conversation from the chat history to review previous findings or continue the analysis from where you stopped.

Chat history helps IT teams maintain continuity across investigations, avoid repeating earlier questions, and build on insights discovered during previous sessions.

Context retention

Within a conversation, Ask Zia interprets each question in relation to the earlier discussion. This lets you ask follow-up questions without repeating the same metrics, filters, or IT context.

  • Type-ahead assistance: Suggests relevant columns, dimensions, measures, functions, visualizations, and existing reports while you frame a question.
  • Follow-up suggestions: Recommends related metrics, filters, and analytical questions based on the current conversation.
  • Proactive recommendations: Suggests useful questions using the data model, previous interactions, and commonly used query patterns.

Best practices

  • Begin with a simple question, and refine it through follow-up questions.
  • Use metric and dimension names that match the columns in the workspace.
  • Include relevant filters such as service, site, priority, category, or time period.
  • Group metrics by useful dimensions when comparing IT services, teams, locations, or infrastructure components.
  • Specify a time period when investigating trends, anomalies, or forecasts.
  • Add detail gradually instead of placing every analytical requirement in the first prompt.

Embed Ask Zia

You can embed Ask Zia in other applications to provide conversational analysis without requiring users to switch applications. Embedded users can explore permitted data, investigate patterns and anomalies, and review the factors influencing important IT metrics. Refer to Embed Ask Zia for details.

AI architecture, security, and governance

Analytics Plus administrators can choose the AI services available for supported tasks and manage the information shared with those services. The information sent for an AI request depends on the task and can include the prompt and metadata such as column names, data types, and the data model. Columns identified as personally identifiable information are not shared. For insight recommendations, aggregated data is used.

Use GenAI Configuration to review and manage the applicable AI settings.

Explore Ask Zia