Using Agent and AI Insights
Agent and AI Insights (powered by the Analyst Agent) lets you ask questions about your data in natural language and get instant, data-driven answers with visualizations.
What is Agent and AI Insights?
Agent and AI Insights is your AI-powered analytics assistant. Instead of navigating through dashboards and building reports, you can simply ask questions like:
- "How did our cart abandonment journey perform last week?"
- "Which treatments have the highest CTR?"
- "What changed since last month?"
The Analyst Agent will analyze your data and provide clear answers with charts and recommendations.
Accessing Agent and AI Insights
- Click Agent and AI Insights in the sidebar
- The chat interface opens
How to Ask Questions
The Chat Interface
The Agent and AI Insights screen has:
- Chat input - Where you type your questions
- Conversation history - Previous questions and answers
- Visualizations - Charts and graphs generated from your data
- Actions - Buttons to explore further or take action
Asking Effective Questions
Be Specific:
❌ "How are things going?"
✅ "What was our conversion rate for the Welcome Journey last week?"
Include Context:
❌ "Compare performance"
✅ "Compare CTR between our Holiday Sale and Summer Sale journeys"
Specify Time Periods:
❌ "Show me conversions"
✅ "Show me daily conversions for the past 30 days"
Example Questions
Performance Analysis
"What were our top 5 performing treatments last month?"
"How does this week's CTR compare to last week?"
"Which journeys drove the most conversions in Q4?"
"What's our overall conversion rate trend over the past 90 days?"
Troubleshooting
"Why did conversions drop on December 15th?"
"Which treatments are underperforming compared to average?"
"Are there any journeys with significantly lower engagement?"
"What changed in our performance since last Monday?"
Optimization Ideas
"Which user segments are most engaged with our treatments?"
"What time of day do we see the highest click-through rates?"
"Which treatment types perform best for new users?"
"Where should we focus to improve our welcome journey?"
Comparisons
"Compare performance between mobile and web users"
"How does Treatment A compare to Treatment B?"
"Which is performing better: our Evergreen or Burst journeys?"
Understanding Responses
Text Answers
The Analyst Agent provides written explanations:
Example Response: "Your Welcome Journey had a 4.2% CTR last week, which is 0.5% higher than the previous week. The 'Day 1 Welcome' treatment was your top performer with a 6.1% CTR."
Charts and Visualizations
Responses often include visual data:
- Line charts - Trends over time
- Bar charts - Comparisons between items
- Tables - Detailed metric breakdowns
Actionable Insights
The agent may suggest actions:
- "Consider pausing Treatment X which has below-average CTR"
- "Your top users respond best to personalized messages"
- "The Holiday journey drove 35% more conversions - consider extending similar journeys"
Taking Action from Insights
When Agent and AI Insights identifies something actionable, you can:
Quick Actions
Some insights include action buttons:
- "View this treatment" - Opens treatment details
- "Go to journey" - Navigates to journey settings
- "See full analytics" - Opens detailed analytics page
Manual Follow-up
Based on insights, you might:
- Pause underperforming treatments
- Increase allocation to top performers
- Create new treatments similar to winners
- Adjust targeting/guardrails
Best Practices
Start Broad, Then Narrow Down
- Start with a general question: "How did we perform last week?"
- Follow up based on the answer: "Tell me more about the drop on Wednesday"
- Drill into specifics: "Which treatments were affected?"
Use Follow-up Questions
The agent remembers context, so you can have a conversation:
You: "What was our CTR last month?"
Agent: "Your overall CTR was 4.5%..."
You: "How does that compare to the previous month?"
Agent: "It's up 0.3% from 4.2%..."
You: "Which treatments drove that improvement?"
Save Important Insights
When you get valuable insights:
- Take screenshots for presentations
- Note key findings for team discussions
- Export data if you need detailed analysis
Regular Check-ins
Make Agent and AI Insights part of your routine:
- Daily: "Any anomalies in yesterday's performance?"
- Weekly: "What were our highlights and lowlights this week?"
- Monthly: "How did we perform against our goals this month?"
What Agent and AI Insights Can and Can't Do
What It's Good For
✅ Analyzing historical performance data ✅ Comparing treatments, journeys, and time periods ✅ Identifying trends and anomalies ✅ Suggesting optimization opportunities ✅ Answering "what happened" questions
Limitations
❌ Cannot make changes for you (you must take action manually) ❌ Cannot predict future performance with certainty ❌ May not have access to all data sources ❌ Complex multi-step analyses may need multiple questions
Note: Agent and AI Insights analyzes data from Auxia. It doesn't have access to external systems unless integrated.
Prompt Templates
Use these templates as starting points:
Performance Review
"Give me a performance summary for [journey name] for the [time period]"
Comparison
"Compare [metric] between [item A] and [item B] for [time period]"
Trend Analysis
"Show me the trend for [metric] over the past [time period]"
Troubleshooting
"Why did [metric] change on [date] for [journey/treatment]?"
Recommendations
"What are the top opportunities to improve [metric] for [journey]?"
Troubleshooting
"The agent doesn't understand my question"
Try:
- Rephrasing with simpler terms
- Being more specific about what you want
- Breaking complex questions into parts
"The data seems wrong"
Check:
- Your date range selection
- Whether you're looking at the right journey/treatment
- If there's a lag in data (can be up to 24 hours)
"I don't see Agent and AI Insights in my menu"
You may not have permission to use this feature. Contact your admin to request access.
Next Steps
Now that you understand Agent and AI Insights:
- Quick Reference - Common shortcuts and tips
- Return to Understanding Analytics for dashboard deep-dives
For advanced Agent and AI Insights usage, see the Comprehensive Manual - Agent and AI Insights.