Introduction
The Engagement Analytics dashboard provides valuable insights into how your AI Concierge initiates and participates in conversations with website visitors. This powerful tool helps you measure visitor interaction rates, understand conversation origins, and track lead collection, allowing you to optimize how your AI converts passive browsers into active shoppers or support seekers.
Accessing the Engagement Dashboard
Log in to your Rep Console
In the left navigation menu, open AI Insights
Select Engagement Analytics
Understanding Your Engagement Metrics
Your Engagement dashboard displays key performance indicators that update based on your selected date range:
Engagement Rate
What it shows: The percentage of total website visitors who engaged in a chatbot conversation.
Why it matters: This metric reveals how effectively your AI attracts visitor attention and initiates meaningful interactions.
Example: 0.03% engagement rate means 3 out of every 10,000 website visitors had a conversation with your AI.
Conversations by Traffic Source
What it shows: A breakdown of AI conversations based on how visitors arrived at your site.
Why it matters: This helps you understand which traffic channels produce visitors most likely to engage with your AI.
Tracked sources include:
Direct (visitors who typed your URL or used a bookmark)
Referral (visitors from links on other websites)
Search (visitors from search engines)
Social (visitors from social media platforms)
Instagram DM (direct messages via Instagram)
Facebook DM (direct messages via Facebook)
Whatsapp DM (direct messages via Whatsapp)
Conversations by Customer Type
What it shows: Segmentation of AI chats by new versus returning customers.
Why it matters: This helps you understand how your AI performs with first-time visitors compared to repeat customers.
Example: 60% new customers, 40% returning customers.
Engagement Rate from Proactive Approach
What it shows: The percentage of previously disengaged visitors who had a conversation after the AI initiated contact.
Why it matters: This measures how effective your proactive outreach settings are at converting passive browsers into active conversations.
Example: 1.50% means that out of all visitors approached by the AI, 1.5% engaged in a conversation.
Frictionless AI Conversations
What it shows: The number of AI conversations that ran without a friction signal, meaning nothing in the conversation suggested the shopper wasn't fully satisfied.
A conversation is not counted if the shopper:
asked something the AI couldn't answer
gave negative feedback
asked to reach your team, even if the AI then helped them and no one from your team ever joined
answered the AI's approach with "No" or "Continue shopping"
Why it matters: This is your cleanest view of how often the AI handles a shopper smoothly from start to finish. A rising count means more shoppers are getting what they need without hitting a snag.
Example: 3,000 Frictionless AI Conversations in the selected period means 3,000 conversations ran without any sign of friction. The tile shows a count, not a percentage.
See the full definition on the tile: Select the ⓘ icon next to the tile's name. It works with a mouse, with a keyboard (Tab, then Enter), or with a tap on mobile.
Why this number differs from "Resolved by AI": "Resolved by AI" in Support Analytics is a different measure, so the two numbers won't match. For example, a shopper who asks for a human but is then helped by the AI can still count as resolved by AI, but isn't counted as a frictionless conversation.
Proactive AI vs. Shopper-Initiated Conversations
What it shows: A comparison of conversations started by your AI versus those initiated by shoppers.
Why it matters: This helps you understand the balance between proactive and reactive engagement strategies.
Example: 70.37% AI-initiated, 29.63% shopper-initiated.
Leads Collected
What it shows: The number of email addresses or phone numbers gathered by your AI during conversations.
Why it matters: This tracks how effectively your AI captures valuable contact information for future marketing or sales follow-up.
Example: 7 leads collected during the selected time period.
Working with Engagement Analytics
Comparing Performance
Toggle the "Compare to previous period" switch to see how your current metrics compare to the previous time period of equal length. This feature helps you identify trends and measure the impact of any changes you've made to your AI configuration.
Filtering Your Data
Click the Filters button to refine your data by:
Customers: All, New, or Returning
Traffic Source: All, Direct, Referral, Search, or Social
Platform: All, Web (Desktop + Mobile), Desktop only, Mobile only, or Simulator
Date Range Selection
Select your preferred time period using:
Preset options like Today, Yesterday, This Week, or Last Month
The calendar interface for custom date ranges
Quick selectors for common time periods
Viewing Detailed Conversations
Each metric card includes a "See conversations" link that allows you to review the actual chat transcripts that contributed to that metric, giving you valuable qualitative context for your quantitative data.
Tips for Maximizing Engagement Analytics
Regular Review: Check your Engagement Analytics weekly to track performance trends and identify opportunities for improvement.
Optimize Proactive Approach: If your proactive engagement rate is low, consider adjusting your AI's conversation starters or trigger conditions in the Sales Skills section.
Traffic Source Optimization: Focus your AI customization efforts on the traffic sources that show the highest engagement rates.
New vs. Returning Balance: If there's a significant difference in how new and returning customers engage with your AI, create targeted experiences for each group.
Test Different Triggers: Experiment with different proactive triggers (time on page, scroll depth, exit intent) to find what works best for your audience.
Lead Collection Strategy: Review conversations where leads were successfully collected to identify effective approaches that can be replicated.
Troubleshooting
Low Engagement Rate: Check that your AI widget is properly configured and visible on your site. Consider more engaging conversation starters or more prominent widget placement.
Incorrect Metric Display with Filters: If applying filters causes metrics to display inconsistent values (especially percentages that don't add up to 100%), try refreshing the page or temporarily disabling filters to view accurate data.
Missing Data: Ensure your tracking is properly implemented. If you've recently installed Rep AI, note that it may take some time to accumulate meaningful data.
Platform-Specific Issues: If you notice significantly different engagement rates across platforms, review your mobile optimization settings.
Frictionless AI Conversations doesn't match "Resolved by AI": This is expected. The two tiles measure different things. Frictionless AI Conversations leaves out any conversation where the shopper asked to reach your team, even if the AI then resolved it, while "Resolved by AI" in Support Analytics can still count that conversation. Dividing Frictionless AI Conversations by another tile's total won't give you the "Resolved by AI" rate, because the two aren't built on the same set of conversations.