Sagent
Find everything you need to set up, customize, and optimize your Sagent. This section guides you through creating an effective and working guided selling Sagent.
Sagent is an AI-powered guided selling chatbot. It uses your product catalog and category knowledge to guide shoppers through natural conversations and recommend relevant products.
Sagent lifecycle
Use this journey to take Sagent from initial setup to continuous improvement.
Complete prerequisites
Install the Crobox Snippet Implementation before deploying Sagent.
Set up and sync the Setting up a Product Feed. Then select the products and locales Sagent can use.
Conversation analytics do not require pageviews. Add Event Tracking Implementation when you need analytics that depend on site activity.
Configure Sagent
Use Setup to configure core settings, content, tools, and design.
Use Data to configure products, knowledge, and guidance. Use Behavior to control questions and recommendations.
Add Translations for every supported language.
Validate before launch
Use Validate before launch to test catalog coverage, shopper conversations, tools, and entry points.
Create configurable response and tool-use checks in Test Cases. Run and investigate them in Test Results.
Monitor live conversations
Use Monitoring to review conversation health, tool activity, and flagged sessions.
Use Analytics to track engagement and investigate conversation data.
Current capability boundaries
Sagent has an accessibility foundation. Accessibility testing is still pending. Do not treat Sagent as compliant with a specific accessibility standard until that testing is complete.
Last updated
Was this helpful?
