Conversational agent integration

Integrating Conversational agent is the first step toward delivering a conversational shopping experience that drives your business goals. This article provides an overview of how to integrate Conversational agent into your website.

Integration timeline

The integration follows these key milestones.

Conversational agent integration

If you are a Marketing customer, you don’t need to add the Bloomreach tracking pixel. Non-Marketing customers must implement it.

First use case launch

Once Conversational agent is set up, you can proceed with launching your first use case.

Conversational agent trial

If you’re exploring Conversational agent before a full integration, use the Conversational agent Chrome extension to test Conversational agent’s capabilities with your website data—without needing to install the Marketing SDK.

Conversational agent integration process

Set up your Conversational agent project

Bloomreach creates a new Marketing project for you and enables the Conversational agent module. This setup typically takes less than 1 business day, allowing you to start the integration right away.

Assign user roles

Once the Marketing project setup is set up, assign permissions under Access management. Define a clear responsibility matrix and implement role-based access control to ensure everyone has the right level of access.

Import product feed

Ensure your product catalog is formatted correctly and contains essential data such as:

• Title
• Absolute URL and image
• Pricing
• All product detail page (PDP) data
• Metadata

The quality of your product feed affects Conversational agent’s performance. Keep your catalog updated to ensure accurate pricing, descriptions, and discount responses.

📘

Note

If you are an existing customer, use a regularly updated catalog to maintain data accuracy.

Set up event tracking

Event tracking helps Conversational agent personalize experiences and measure impact.

Minimum events to track:

• purchase
• view_item
• cart_update

See the Conversational agent event tracking article for a full list of required and recommended events.

📘

Tip

Use the Exponea Console to explore event tracking. For more details, review the Exponea documentation or install the Exponea plug-in.

Conversational agent stores user interaction events. To minimize risks associated with storing personally identifiable information (PII), follow the data storage guidelines.

Create an agent

Create and configure an agent to achieve a specific business goal, such as increasing conversions.

Configure Conversational agent

Optimize Conversational agent’s performance by configuring these settings in the back office:

General knowledge data

• Provide merchant information (merchant_info), including a short description of your business and main product categories.
  • Note: Sending merchant information in English is preferable, but if your merchant information isn't available in English, you may send it in the local language set for the agent.
• Upload fresh data sources to improve response accuracy.

Language, currency, and tone of voice

Align Conversational agent with your brand’s communication style.

Define a use case

Select a use case from our extensive use case suite or create a custom one tailored to your business objectives:

• Goal: What do you want to achieve? For example, increase add-to-cart actions.
• Audience: Who should see Conversational agent? For example, indecisive shoppers who haven't added anything to their cart in 2 days.
• Timing: When should Conversational agent engage shoppers? For example, after a customer views 3 products without taking action.
• Engagement: How should Conversational agent interact with shoppers? For example, offering product recommendations.

You can run multiple use cases at the same time.

Set up weblayers

Conversational agent uses Marketing weblayers for advanced visual customization and user targeting. Configure the following:

• When and where Conversational agent appears: Specify locations (for example, mobile or pc) and pages (PDP, checkout) where Conversational agent should engage shoppers.
• Dismissal settings: Set how long Conversational agent remains hidden after being closed.
• Audience segmentation: Ensure Conversational agent targets the right users.
• Interaction limits: Control how often users see Conversational agent.
• A/B testing: Set up A/B tests to measure impact.
• Design customization: Align visuals and messaging with your brand’s guidelines.
• Display hours: Schedule when Conversational agent should be active and hide conflicting elements on the page.

📘

Note

A Bloomreach Solutions Consultant will assist with the technical setup. Your role is to define the use case and key configurations.

Expand to more use cases

Once your first use case is live, consider expanding by configuring additional use cases:

• Monitor performance: Use the Conversational agent evaluation dashboards to track impact and identify improvements.
• Configure new use cases: Based on performance insights, select and configure additional use cases.

Initiatives in your project

Initiatives are core components that support specific engagement strategies in Conversational agent. Each initiative optimizes your Conversational agent project setup.

• Conversational agent core initiative: Manages follow-up weblayers that display the Conversational agent chat window to users who’ve already interacted with Conversational agent.
• Conversational agent emergency initiative: Handles emergency scenarios and manages error notification mailing list.
• Use case initiatives: Each use case has its own initiative (for example, PLP or PDP use cases). The use case initiative includes the trigger tag that starts or stops the initial Conversational agent trigger.

These initiatives help Conversational agent engage the right users at the right time.

Bloomreach Academy

Conversational agent integrates with Marketing, which powers its analytics and data management capabilities. To get the most out of this integration, we recommend these courses:

• Engagement Crash Course: Understand how to store and use rich customer data.
• Engagement Advanced Analytics: Build in-depth custom evaluations of Conversational agent’s performance and gain actionable insights.
• Engagement Integration: A technical course for users integrating Marketing and building custom use cases.

Support and feature requests

If updating your data doesn’t improve conversation quality or you encounter issues:

• Verify your product catalog and event tracking setup.
• Contact your CSM to report bugs or request new features.