Agentic Analytics

Agentic Analytics is an AI-powered analytics experience for CX leaders. It brings your contact center data into a single view, proactively highlights what needs attention, and lets you ask questions about your operation in plain language — no reports to build and no coding required.

With Agentic Analytics you can:

• Start each day with a proactive briefing of what needs your attention.

• Review active signals surfaced from your operations data.

• See prioritized recommendations with the potential impact of acting on them.

• Ask questions about your data in natural language.

Navigation

Agentic Analytics is organized through the left navigation:

pages	What it contains
Home	Your daily briefing, active signals, top recommended actions, and the Ask AI panel.
Automation Opportunities	Interactions that can be automated by an AI agent, with projected impact.
AI Agents	AI agents and how they are performing.
Action History	A record of actions you have taken.
Your Saved Dashboards	Views you have saved for repeated use.
Recent	Your recent conversations.

Home

The Home page is the starting point in Agentic Analytics. It gives CX leaders a single view of what needs attention across the operation, without having to open a separate report. The Home page includes following sections:

sections	What it contains
Your daily briefing	A proactive narrative summarizing what was found – including anomalies in your data and automation opportunities worth implementing. A summary row highlights the number of Signals, Recommended actions, Projected annual savings, and Interactions deflectable per year.
Signals	Active signals surfaced from your operations data, updated every 24 hours. Each signal card describes the finding and shows a severity label.
Top Recommended Actions	Prioritized recommended actions. Each card shows the potential interaction savings, category of action type (for example, Automation Opportunity), projected annual savings, and a link to act on the recommendation.
Ask AI	A conversational panel for asking questions about your data in plain language, with suggested questions to help you get started.

Ask AI

The Ask AI panel sits at the bottom of the Home page and lets you ask questions about your data in plain language — no report building or query syntax required. Instead of navigating to a specific report, you describe what you want to know and Agentic Analytics interprets the question against your operations data.

Asking a Question

You can ask a question in two ways:

• Type your question into the panel and click Send.

• Click one of the suggested questions above the panel to get started.

Suggested questions give you a quick way in without having to phrase a question yourself. Examples include:

• What are my AI agents currently struggling with?

• What's causing repeat calls?

• How can I improve my CSAT A standard call center metric used to measure the level of customer satisfaction with a product, service, or interaction, typically captured immediately after the customer’s experience.?

• What tasks are my human agents spending extra time on?

What You Can Ask

Ask AI is designed for the kinds of operational questions a CX leader asks every day, including:

• Performance: How a metric, team, skill, or agent is trending.

• Root cause: Why a metric is moving or what is driving repeat contacts.

• Opportunities: Where interactions could be automated or where there is room to improve.

• Your hybrid workforce: Questions about both your human agents and your AI agents.

Because questions are interpreted against your unified data, you can phrase them naturally and follow up to go deeper.

Your Recent Questions

Questions you have asked appear under Recent in the left navigation, so you can return to them without retyping. Click View All to see your full history. You can also save views you return to often under Saved Dashboards.

Automation Opportunities

Automation Opportunities shows you which customer intents in your contact center are worth automating, how much automating them could save you, and lets you create and publish an AI agent to act on them — without leaving the page. With Automation Opportunities you can:

• See which categories and topics have the highest automation potential, ranked by projected annual savings.

• Review the volumes, sentiment, and top agent actions behind each opportunity.

• Create an AI agent for a category or topic directly from its card.

• View, edit, and publish an AI agent, and track its status from Draft to Published.

• Ask questions about your automation opportunities in plain language.

Navigation

Automation Opportunities has two tabs, Categories and Topics, scoping the page to that level of your intent hierarchy. A filter bar above the tabs lets you narrow the page by Date Range, Team, Skill, Channel, Category, and Direction.

sections	What it contains
AI briefing	A short AI-generated summary of the highest-confidence opportunities for your current tab and filters.
Analyzed data	Total Interactions, Avg. Duration, Avg. Sentiment, Intents Identified, Actions Identified, and Agents for the selected period.
Top Opportunities	Ranked cards for the active tab, each showing volume, automatable volume, and projected annual savings. Click Explore Breakdown & Related Opportunities to see the top agent actions and related topics underneath a category.
Ask	Opens the AI Assistant panel to ask questions about your automation opportunities.

Viewing Sample Interactions

Click the three-dot menu on any category or topic card, then click View Sample Interactions to see real interaction examples behind that opportunity. You can filter by Date, Channel, Team, and Skill, and click the play icon to review the recording.

Create AI Agent

Click Create AI Agent on a category or topic card to have Agentic Analytics build an agent for that intent. The AI Assistant panel opens and works through the request, then confirms the agent is ready with a View Agent button. On the card, Create AI Agent changes to View Agent.

View AI Agent

Click View Agent to open the agent’s detail page, where you can review or edit it in two views:

• Flow View: A visual map of the agent and the skills and tools it can invoke.

• Textual View: An editable form covering the agent’s overview and each of its skills along with its tools.

Publish AI Agent

Click Publish AI Agent on the agent’s detail page to publish it to the AI Agents platform. Its status changes from Draft to Published, and you can access the published agent through the Open Published AI Agent link.

Ask AI in Automation Opportunities

Click Ask in the top-right corner to open the AI Assistant without leaving Automation Opportunities. The suggested questions above the panel help you get started — “What are the key insights here?”, “What should I focus on first?”, “What changed recently?”, “What actions do you recommend?” — or type your own question about a specific category or topic, such as:

• “What’s the automation opportunity for Bill Explanation?”

• “Which topics have the highest automatable volume this month?”

• “Why is the automatable % for Fares & Cancellations so low?”

• “Create an AI Agent for Flight Booking Management.”

If you do not name a category or topic, the assistant defaults to topic level. Answers are grounded in your current tab and filters.

AI Agents

AI Agents shows you how your AI Agents (Cognigy) are performing, where they are falling short, and gives you what you need to act on improvement recommendations directly. With AI Agents you can:

• See how each agent is performing against evaluation criteria.

• Review flagged sessions that may need attention.

• See AI-generated recommendations for improving agent configuration.

• Apply approved fixes directly to your agent in AI Agents (Cognigy).

Navigation

AI Agents is organized across four tabs — Overview, Goals & Outcomes, Guardian AI Scorecard, and Intent & NLU. A filter bar above the tabs lets you narrow the page by Date Range, Project, and other dimensions.

sections	What it contains
Overview	A summary of your AI agents and their current performance.
Goals & Outcomes	How your AI agents are performing against the goals you have set.
Guardian AI Scorecard	Evaluation scores by criterion, flagged sessions, and AI-generated improvement recommendations.
Intent & NLU	How well your agents are understanding and handling customer intents.

Guardian AI Scorecard

The Guardian AI Scorecard uses AI to evaluate your agent's conversation transcripts against a set of criteria — system-defined or ones you define yourself. For each session, the AI acts as a judge, reviewing the transcript and classifying it against each criterion. Those classifications build up over time into trends and aggregate scores that show you how your agent is performing and where it is consistently falling short.

The results from the Guardian AI Scorecard provides the foundation for Improve AI Agents. It identifies recurring gaps by analyzing criteria that sessions fail at statistically significant rates, then uses those patterns to generate targeted recommendations for improvement.

sections	What it contains
AI Insights	A summary of the most significant findings for the current period, including the top gaps identified and links to act on recommendations.
AI Agent Optimization	The top gap patterns identified from evaluation results, with a link to review and apply fixes in Improve AI Agents.
Key Performance Indicators	High-level scores for the current period, including overall pass rate, brand alignment, and compliance.
System Criteria	Evaluation results for each built-in criterion — including pass rate, classification distribution, and trend over time. Compliance issues are flagged with a warning.
User-Defined Criteria	Evaluation results for any custom criteria you have configured.
Sessions to Investigate	Individual sessions flagged for review, with classification outcomes per criterion and links to session transcripts.