> ## Documentation Index
> Fetch the complete documentation index at: https://help.gostanna.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Lars - AI Assistant
> Your AI-powered assistant for customer success automation
# Meet Lars
Lars is Stanna's AI assistant that helps you manage customer success more effectively. Ask questions in plain English to analyze client health, identify risks, and get actionable recommendations.
## Key Features
Ask questions in plain English. Lars understands context and provides conversational, actionable responses
Creates tasks, sends notifications, updates records, and generates insights automatically
Identifies at-risk clients before they churn by analyzing sentiment, engagement, and interaction patterns
Accesses complete client history including interactions, sentiment, health scores, and business metrics
## What You Can Ask Lars
Lars can help with a wide range of customer success tasks:
* "Which clients are at risk this month?"
* "Show me sentiment trends for TechFlow Inc"
* "What's the health score for DataCorp?"
* "Who hasn't been contacted in the last 2 weeks?"
* "Create a follow-up task for all at-risk clients"
* "Schedule check-ins for clients with declining sentiment"
* "Remind me to contact TechFlow next week"
* "Add a note to the CloudTech account"
* "Summarize last week's client interactions"
* "What are the common themes in negative feedback?"
* "Show me MRR at risk this quarter"
* "Which clients have improved the most?"
* "What should I prioritize today?"
* "How can I improve engagement with at-risk clients?"
* "Which clients are expansion opportunities?"
* "What patterns do my churned clients have?"
## How Lars Works
Lars uses advanced AI to understand your needs and provide intelligent assistance:
### 1. Natural Language Processing
* **Understands context**: Recognizes what you're asking even with casual language
* **Intent detection**: Identifies whether you want analysis, actions, or insights
* **Multi-turn conversations**: Remembers previous messages in your conversation
* **Smart clarification**: Asks follow-up questions when needed
### 2. Real-Time Data Analysis
* **Client health scoring**: Analyzes current and historical health metrics
* **Sentiment analysis**: Reviews communication patterns and trends
* **Engagement tracking**: Monitors interaction frequency and quality
* **Risk prediction**: Identifies patterns that indicate churn risk
### 3. Intelligent Recommendations
Lars doesn't just show data—it provides actionable insights:
* **Prioritized actions**: Suggests what to work on first
* **Context-aware advice**: Recommendations based on your full client history
* **Pattern recognition**: Identifies trends across your portfolio
* **Success strategies**: Suggests approaches that worked for similar clients
### 4. Automated Actions
With your approval, Lars can:
* ✅ Create and assign tasks
* ✅ Send email notifications
* ✅ Update client records
* ✅ Generate reports
* ✅ Schedule follow-ups
* ✅ Add notes and attributions
## How to Use Lars
### Access Lars
Find the chat widget in the bottom-right corner of any page in Stanna. Click the sparkle icon to open or press `⌘L` (Cmd+L).
### Ask Questions
Type your question in plain English:
* "Which clients are at risk this month?"
* "Show me sentiment trends for TechFlow Inc"
* "Create follow-up tasks for clients I haven't contacted in 2 weeks"
* "What should I prioritize today?"
### Review Suggestions
Lars provides actionable recommendations with each response. Review suggested actions before executing them.
### Execute Actions
Click "Execute" on any suggested action. Actions are categorized by risk level:
* **Low Risk**: Generate reports, create tasks (can auto-execute)
* **Medium Risk**: Update scores, schedule meetings (requires approval)
* **High Risk**: Send emails, modify client data (requires confirmation)
## Configuration
Configure Lars behavior in Settings:
* **Auto-execute low-risk actions**: Enable for faster workflow
* **Response style**: Choose brief, detailed, or technical responses
* **Risk threshold**: Set sensitivity for risk detection
## Best Practices
### Writing Effective Queries
Be specific and conversational. Lars understands natural language, so ask questions like you would to a colleague.
**Effective Examples:**
* "Which clients are at risk this month and why?"
* "Create tasks to follow up with clients I haven't contacted in 2 weeks"
* "Show me sentiment trends for my Enterprise clients"
* "What should I focus on today based on health scores?"
**Less Effective Examples:**
* "Help" (too vague)
* "Clients" (no clear intent)
* "Do something" (no specific action)
### Getting the Most from Lars
Begin with straightforward questions like "Show at-risk clients" to get familiar with Lars
Add details like timeframes, client names, or specific metrics for more precise results
Always review Lars's recommendations before approving automated actions
Use thumbs up/down reactions to help Lars learn and improve responses over time
### Privacy and Security
Lars processes data securely within your workspace. No client information is shared with external AI services or third parties.
* **Workspace isolation**: Only accesses data within your workspace
* **Permission-based**: Respects user roles and permissions
* **Audit logging**: All actions are logged for compliance
* **Data encryption**: Data is encrypted in transit and at rest
## Common Use Cases
"What clients need my attention this week?" Lars provides a prioritized list with context
"Show me all renewals in the next 30 days and their health scores" Get a comprehensive renewal report
"Create urgent follow-up tasks for all clients with health below 35" Automate your risk response
"Summarize my portfolio health changes this month" Get executive-ready insights
## Troubleshooting
**Quick fixes:**
* Refresh your browser
* Check your internet connection
* Clear browser cache and reload
* Try the keyboard shortcut `⌘/` to reopen
**If still not working:**
Contact support through Settings or email [support@gostanna.com](mailto:support@gostanna.com) with details about when the issue started
**Improve response quality:**
* Be more specific in your questions
* Include timeframes (e.g., "last 30 days")
* Mention specific clients by name
* Break complex questions into smaller parts
**Example of better phrasing:**
Instead of: "Show clients"
Try: "Show me clients with health scores below 40 in the last week"
**Check these common causes:**
* Did you approve the action? (Look for the "Execute" button)
* Do you have permission for this action type?
* Is the client/workspace relationship valid?
**Review action status:**
Click on any suggested action to see its current state and any error messages
**Try rephrasing:**
* Use simpler, more direct language
* Break compound questions into separate queries
* Provide more context about what you need
**Example:**
Instead of: "What's going on with my stuff?"
Try: "Which of my clients have declining health scores?"
## Next Steps
Tips for getting the most out of Lars
Learn about client management features