Accessing the AI Analyst
Navigate to Intelligence → AI Analyst from the main menu. You’ll see the AI chat interface with:- Chat history panel (center)
- Input box at bottom
- Suggested prompts when starting a new conversation
- Clear conversation button (refresh icon, top right)
What the AI Analyst Can Do
The AI has complete access to your business data:Data Sources Analyzed
- Appointments: All 100+ appointments across 5 locations
- Guests: 50 client records with visit history, CLV, preferences
- Invoices: Revenue by service, provider, location
- Daily Metrics: 90 days of performance data (450 total records)
- Conversations: 50 AI interactions with sentiment analysis
- Alerts: 20 active opportunities and warnings
- Agent Performance: Status of 9 AI agents across modules
Types of Questions You Can Ask
Performance Analysis- “Which center is underperforming this month?”
- “How much revenue did we generate last week?”
- “What’s our average response time?”
- “Compare SoHo vs Williamsburg performance”
- “Which location has the best utilization rate?”
- “Show me no-show rates by location”
- “Who are our top 10 clients by CLV?”
- “Which clients are at risk of churning?”
- “How many new clients did we acquire this month?”
- “Which service generates the most revenue?”
- “What’s the average booking rate for Body Contouring?”
- “Compare Botox vs Filler performance”
- “Where are the biggest revenue gaps right now?”
- “How much revenue are we losing to no-shows?”
- “What should I focus on this week?”
- “Which provider has the best rebooking rate?”
- “How many appointments did Dr. Chen complete this week?”
- “What’s the utilization rate by provider?”
Suggested Prompts
When you first open the AI Analyst, you’ll see 6 suggested prompts:-
“Which center is underperforming this month?”
- Analyzes revenue trends across locations
- Identifies locations below targets
- Suggests focus areas
-
“How much revenue are we losing to no-shows?”
- Calculates missed revenue from no-show appointments
- Shows improvement since AI implementation
- Recommends prevention strategies
-
“Compare SoHo vs Williamsburg performance”
- Side-by-side metrics comparison
- Revenue, utilization, no-shows, new clients
- Relative strengths and weaknesses
-
“What should I focus on this week?”
- Prioritized action items
- Based on current alerts and opportunities
- Considers revenue impact and urgency
-
“Where are the biggest revenue gaps right now?”
- Identifies underutilized capacity
- Highlights unfilled time slots
- Suggests promotional opportunities
-
“Which provider has the best rebooking rate?”
- Ranks providers by client retention
- Shows best practices to replicate
- Identifies coaching opportunities
These prompts are visible both at the start of a conversation and as quick-access buttons during your chat session.
Understanding AI Responses
The AI formats responses using markdown for clarity:Response Components
Summary Section- Quick answer to your question
- Key findings highlighted
- Usually 2-3 sentences
- Deeper breakdown with numbers
- Comparisons to benchmarks or trends
- Context explaining what the numbers mean
- Formatted tables for multi-item comparisons
- Sorted by relevance (highest to lowest impact)
- Columns include metrics and trends
- 3-5 actionable next steps
- Prioritized by impact
- Specific, not generic advice
- Bold text: Key metrics and findings
- Italics: Context and explanations
- Bullet lists: Multiple related items
- Numbered lists: Sequential action steps
- Tables: Comparative data
- Code blocks: Specific values or formulas
Example Response Format
Location-Specific Queries
The AI automatically filters data based on your currently selected location in the top navigation.How Location Filtering Works
When “All Centers” is selected:- AI analyzes aggregated data across all 5 locations
- Provides network-wide insights
- Can break down by individual location within the response
- AI focuses only on that location’s data
- Comparisons are vs. that location’s historical performance
- Can still reference other locations for context if asked
Changing Location Mid-Conversation
If you change the location selector while chatting:- Future questions use the new location filter
- Previous chat messages remain unchanged
- You’ll see a note in the interface indicating the data source
Complex Analysis Queries
The AI can handle sophisticated multi-part questions:Time-Based Analysis
Examples:- “Show me revenue trends for the last 3 months”
- “Compare this week vs. last week performance”
- “What days of the week perform best?”
- Analyzes 90 days of historical metrics
- Identifies seasonal patterns
- Projects future trends based on patterns
Cohort Analysis
Examples:- “How do new clients compare to returning clients in CLV?”
- “What’s the rebooking rate for clients who booked via AI vs. phone?”
- “Compare performance of Injectable services vs. Facial services”
- Segments clients by behavior, service type, booking channel
- Calculates comparative metrics
- Identifies statistically significant differences
Root Cause Analysis
Examples:- “Why is the no-show rate high at Williamsburg?”
- “What’s causing low utilization on Tuesdays?”
- “Why did revenue drop last week at Hoboken?”
- Examines multiple potential factors
- Tests hypotheses against data
- Provides evidence-based explanations
Predictive Questions
Examples:- “What will revenue be if we maintain current growth?”
- “How much could we earn by reducing no-shows to 10%?”
- “What happens if we add one more provider at SoHo?”
- Uses historical trends to project forward
- Models scenarios based on variable changes
- Calculates potential ROI of improvements
Best Practices for Getting Great Answers
Be Specific
❌ Vague: “How are we doing?” ✅ Specific: “What’s our revenue growth rate this month vs. last month?” ❌ Vague: “Tell me about clients” ✅ Specific: “Which clients with CLV over $5K haven’t visited in 60+ days?”Ask Follow-Up Questions
The AI remembers context within a conversation: Initial: “Which location has the highest revenue?” Follow-up: “What’s their utilization rate?” Follow-up: “How does that compare to Williamsburg?” Each question builds on the previous context.Request Specific Formats
You can ask for data in specific formats:- “Show me a table of revenue by location”
- “List the top 5 services by booking count”
- “Give me a bullet list of action items”
- “Rank providers by performance”
Combine Multiple Metrics
Examples of Combined Queries:- “Show me locations ranked by revenue, utilization, and no-show rate”
- “Which clients have high CLV but low visit frequency?”
- “What services have high demand but low provider availability?”
The AI can analyze up to 5-7 different metrics in a single query. More than that and the response becomes too complex.
Understanding AI Context
The AI Analyst has access to computed context that includes:Real-Time Metrics
- Current day bookings
- Today’s revenue so far
- Active conversations
- Pending escalations
Historical Trends
- Last 30 days aggregated metrics
- Previous 30 days for comparison
- 90-day trend data
- Phase-based performance (Pre-AI, Ramp-up, Full Operation)
Client Intelligence
- Individual client history
- CLV calculations
- Visit patterns and preferences
- No-show risk scores
Opportunity Detection
- Active alerts (20 opportunities/warnings)
- Revenue gap identification
- Underutilized capacity
- At-risk high-value clients
Chat Management
Clearing Conversations
Click the refresh icon (top right) to:- Clear current chat history
- Start fresh with new context
- Reset conversation memory
- Switching to a completely different topic
- AI seems confused by previous context
- You want to verify a previous answer independently
Conversation History
The chat remembers:- All questions you asked in current session
- All AI responses
- Follow-up context
- When you navigate away from AI Analyst page
- When you refresh your browser
- Between different user sessions
Streaming Responses
AI responses appear gradually (streaming) rather than all at once:Why Streaming?
- Immediate feedback that AI is working
- Can start reading while response generates
- More natural conversation feel
- Shows complex analysis in progress
While AI is Thinking
You’ll see:- Animated dots in a message bubble
- Brain icon with pulse animation
- Input box is disabled (can’t send new messages)
- Simple queries: 2-3 seconds for complete response
- Complex analysis: 5-8 seconds
- Multi-part questions: 8-12 seconds
Data Accuracy & Sources
All AI responses are based on: Primary Source: Zenoti API- Live appointment data
- Real client records
- Actual invoice amounts
- Provider schedules
- AI conversation analytics
- Sentiment analysis
- Predictive scoring
- Opportunity detection
- Metrics update in real-time as appointments complete
- Daily metrics compiled at midnight
- Trends calculated on-demand when queried
The AI always cites its data source in responses: “Based on Zenoti data from the last 30 days…” This transparency ensures you can verify findings.
Example Conversations
Performance Review
You: “How did we perform last month?” AI: Provides overall revenue, bookings, key trends vs. previous month You: “Which location grew the most?” AI: Identifies top performer with percentage growth and dollar amounts You: “What drove that growth?” AI: Analyzes factors like new clients, service mix, utilization changesProblem Diagnosis
You: “Why is utilization low at Stamford?” AI: Shows utilization rate, compares to other locations, identifies gap You: “What time slots are empty?” AI: Breaks down by day of week and time of day You: “How can I fill those slots?” AI: Recommends targeted promotions, rebooking campaigns, schedule adjustmentsStrategic Planning
You: “What should I focus on this quarter?” AI: Prioritized list of opportunities with revenue impact You: “Tell me more about the no-show opportunity” AI: Detailed breakdown of no-show costs, prevention strategies, expected ROI You: “What if we reduced no-shows by 5 percentage points?” AI: Calculates revenue impact across all locationsLimitations & Boundaries
The AI Analyst is powerful but has boundaries:What AI CANNOT Do
❌ Make actual changes to appointments or settings ❌ Send messages or emails to clients ❌ Access individual conversation details (privacy protected) ❌ Predict specific client behavior with certainty ❌ Provide medical or clinical advice ❌ Access data outside Etienne/Zenoti systemsWhat AI CAN Do
✅ Analyze all historical and current business data ✅ Identify patterns and trends ✅ Recommend specific actions based on data ✅ Compare performance across locations/services/time ✅ Project outcomes based on historical trends ✅ Answer “what if” scenario questionsPrivacy & Security
- AI Analyst uses Claude (Anthropic) with enterprise security
- All data transmitted is encrypted
- No client PII is stored in AI conversation logs
- Conversations are not used to train AI models
- Access is role-based (owner and authorized staff only)
While the AI can see client names and service history for analysis, it doesn’t display sensitive information like payment details or medical records in responses.
Tips for Power Users
Build a Weekly Review Routine
Monday Morning:- “What happened last week across all centers?”
- “Which alerts should I prioritize this week?”
- “Are there any high-CLV clients I should contact?”
- “How are we tracking vs. last week?”
- “Any concerning trends in the last 3 days?”
- “Summarize this week’s performance”
- “What should I focus on next week?”
Compare Before/After Campaign Results
Before launching a promotion: “What’s our current booking rate for Hydrafacials on Tuesdays?” After the campaign: “Compare this week’s Tuesday Hydrafacial bookings to last month’s average”Track Improvement Over Time
Set a baseline: “What’s our current no-show rate at White Plains?” Check progress monthly: “Compare White Plains no-show rate this month vs. 3 months ago”Use AI for Team Meetings
Prep meeting agenda: “What are the top 3 issues to discuss in tomorrow’s team meeting?” Support decisions with data: “Pull the data supporting our decision to add a Saturday provider”Next Steps
- Review Understanding Analytics to learn what metrics mean
- Check the Owner View Guide to see how insights appear on your dashboard
- Practice with suggested prompts to get comfortable with the AI
- Build your weekly AI Analyst routine