Access Insights Assistant

Access to the Insights Assistant is governed strictly by administrative configuration and role-based permissions.

To activate and manage how users interact with Insights Assistant, do the following:

• Feature flag enablement — Turn on the Insights Assistant feature flag globally for the tenant environment.

• Grant permissions — Navigate to the role configuration in Admin Suite and turn on the Access Insights Assistant reporting permission. For more information, see Reporting permissions.

• Set the parameters and reporting properties in Shared Configuration — Navigate to Reporting> Shared Configuration screen in Admin Suite to create role-level default reporting configurations.

  For more information, see Shared Configuration in Configure Insights Assistant.

• Role-based access — The Insights Assistant user interface dynamically adjusts based on the user's active role context. If a user possesses multiple roles, they will only see the conversation histories and data environments tied to their currently active session role to avoid context-bleeding or security risks. For more information on how to grant role-based access, see Role-based access.

Launch Insights Assistant

To launch Insights Assistant, click the Application switcher in the Medallia Web reporting toolbar, and then select Insights Assistant:

Once you access the Insights Assistant screen, you can begin asking questions to your assistant:

1. Message bar — Type your prompt, question, or request. Then, click the arrow to send your message. For more information, see Ask questions.

2. Pro Tips — Lists best practices when asking questions.

3. History — Opens a collapsible side panel that provides immediate access to the Chat History and past search items. For more information, see Access Conversation History

The managing activities include:

• Asking business questions

• Interacting with answers

• Accessing source data

• Accessing and managing Chat History

• Start a new chat

• Provide feedback

• Exporting results

For more information, see Using Insights Assistant.

Supported use cases

Insights Assistant supports a wide range of analytical use cases, allowing users to transition seamlessly from high-level performance trends to highly targeted, localized root causes using natural language.

KPI Deep Dive & Driver Analysis

Diagnosing the underlying factors, topics, or root causes driving key performance metrics — such as Overall Satisfaction (OSAT), Likelihood to Recommend (LTR), or Client Effort — up or down.

Sample Prompts:

• "What are the top five factors impacting our primary satisfaction scores and likelihood to recommend?"

• "Which text analytics topic has the largest impact score on our overall performance metric?"

• "What is the primary root cause behind our high-priority customer alerts?"

Trend Identification & Prioritization

Analyzing specific time frames across all active channels (surveys, transcripts, digital) to surface the largest thematic shifts, mounting operational issues, or areas requiring focused coaching.

Sample Prompts:

• "What emerging trends in the last month require immediate focus and action this month?"

• "What are the core metric and sentiment trends across our digital touchpoints between Q1 and Q2?"

• "For the past quarter, what are the key operational differences in trends and mentions regarding wait/response times?"

Unit & Segment Performance Comparison

Evaluating organizational performance and identifying statistical outliers across geographic hierarchies, organizational divisions, demographic segments, or specific business locations.

Sample Prompts:

• "Which location within our main segment showed the most improved frontline score over the last month?"

• "What are the top themes and sentiment scores when broken down by business division?"

• "What is the lowest-performing operating unit or segment for new customer acquisition?"

Cross-Channel Event & Keyword Monitoring

Querying the AI engine to extract specific verbatim comments, identify sudden anomalies, or track external events across diverse data sources.

Sample Prompts:

• "Are there any mentions or anomalies in our unstructured data regarding critical/emergency situations?"

• "Can you provide specific cross-channel examples of users experiencing application or workflow discrepancies?"

• "Compare the volume and sentiment of competitor mentions between January and April."

Generating Customer Suggestions & Ideation

Surfacing actionable customer suggestions from unstructured text data across all customer touchpoints to guide product, service, or broader operational improvements.

Sample Prompts:

• "What are the top customer suggestions and recommendations regarding our core product features or pricing structures?"

• "What specific feedback trends are coming from our highest-value customer segment?"

• "What core themes are driving positive feedback from our most loyal promoters?"

Data Dictionary & Conversational Explanations

Utilizing the conversational chat interface to instantly explain complex backend metric calculations, active data filters, data structures, or reporting terminology.

Sample Prompts:

• "Explain how the different customer segments in this view are defined."

• "Explain the active filters applied to this segment analysis."

• "Explain the calculation behind this specific impact score and negative sentiment percentage."

Frontline Employee Coaching

Empowering team managers to leverage the Insights Assistant to analyze individual or team performance trends to isolate targeted coaching and improvement opportunities.

Sample Prompts:

• "Which team members or frontline associates showed the most metric improvement over the last week?"

• "What specific customer experience drivers or themes is [Employee Name] currently struggling with or trending downward on?"

• "What are the top coaching opportunities for [Division] in the past 3 months?"

Turning Insights into Action

Generating actionable recommendations, executive communication drafts, and concrete next steps in seconds to help structure and share discovered insights.

Sample Prompts:

• "Draft an email that I can send to leadership with these findings and recommendations."

• "Build an agenda for my next team onsite meeting with key coaching themes by department."

• Command: "Export conversations and insights directly to PDF for real-time sharing."

Insights Assistant Agent

The Insights Assistant agent refers to the underlying generative AI framework or engine that interprets user prompts. The system uses it to perform the following backend tasks:

• Intent Interpretation — It translates a user's natural language question into concrete data filters, parameters, and structural constraints.

• Query Execution — It instructs the AI agent on how to call specific Medallia reporting tools (like the TA Snapshot Tool or Entity Search Tool) to query the data.

• Response Generation — It synthesizes the exact numbers, calculations, and tables returned by the system into cohesive, text-based summary answers.

• Suggested actions — Where applicable, it generates suggested recommendations and actions based on the contents of customer experience data.

• Thread Title Generation — It analyzes the initial user query to automatically generate a concise, human-readable title for the saved thread in the Conversation History.

Essentially, the Insights Assistant agent functions as the primary cognitive layer that bridges natural language conversation with Medallia, Inc.'s structured data calculations.