Skip to main content

AI Product Finder Configuration

This article walks you through the full configuration of the AI Product Finder — a powerful AI agent that helps your end users discover the right products through a conversational WhatsApp experience.

Updated this week

Before You Begin

Make sure the following prerequisites are completed before starting the AI Product Finder configuration:

⚠️ PrerequisitesAI General Configuration: Ensure you have completed the general AI setup, including configuring your AI Agents and defining your AI Persona. Learn more → LINK • AI Product Finder Knowledge Base: Ensure your Product Finder Knowledge Base has been uploaded before proceeding. Without it, the agent cannot search or recommend products.

  • AI General Configuration: Ensure you have completed the general AI setup, including configuring your AI Agents and defining your AI Persona. Learn more → LINK

  • AI Product Finder Knowledge Base: Ensure your Product Finder Knowledge Base has been uploaded before proceeding. Without it, the agent cannot search or recommend products.


Step 1: Create Your AI Product Finder Agent

The first step is to create a dedicated AI agent of the Product Finder type and connect it to your Knowledge Base.

To create your agent:

  1. Navigate to Charles AI Squad → Agents in the left-hand menu.

  2. Click "Create AI Agent".

  3. Select the agent type: Product Finder — this creates a virtual agent designed to help users find their desired products.

  4. Name your agent — choose a clear, recognizable name, such as:

    • "Your Brand Name Product Finder"

    • "Your Brand Name FAQ Agent"

💡 Tip: A descriptive name makes it easier to manage multiple agents within your AI Squad.


Step 2: Define Your AI Product Finder General Settings

Once your agent is created, configure its general behavior settings.

Memory Control

Define how many past messages your AI agent should remember when generating responses. A higher value improves conversation continuity but may slightly increase response time.

  • Range: Up to the last 10 messages

  • Recommendation: Start with a moderate value and adjust based on your use case.

Additional Instructions

You can provide custom guidance to your agent that goes beyond the platform's built-in settings. Use this field to fine-tune how the agent behaves in specific scenarios.

Example instructions:

Price Query: If the user asks for the price ALWAYS guide users to click the 'Jetzt entdecken' button for complete product information, including the price.  Product Ranking in `show_items` (Crucial): When choosing the best results, you MUST ALWAYS choose from five different brands. You MUST NOT choose products from the same brand unless the user specifies the brand.

⚠️ Important: Instructions entered here can override your global AI settings. For example, if emojis are disabled globally but you enable them in this field, the agent-level instruction will take precedence.


Step 3: Configure Your Product Finder Attributes

This step maps your product catalogue fields to the AI agent so it can search and filter products accurately.

3.1 Select Your Product Catalogue

Choose the source of your product data:

  • CSV Upload, or

  • Shopify Catalogue

3.2 Map Fields as Filters & Attributes

In the Agent Configuration, map all relevant product fields as filters and attributes. For each field, provide a clear description of what the attribute represents — this helps the AI understand how to use it when filtering products and making recommendations.

💡 Example: If you have a field called color, describe it as "The color of the product, used to filter results when a user specifies a preferred color."

3.3 Define Mandatory Fields

Mark specific fields as mandatory if the AI agent should always ask the user for that information before performing a search — unless the user has already provided it in the conversation.

3.4 Boost or Bury Results

You can influence the ranking of product results by defining Boost or Bury rules for specific attributes. This allows you to promote certain products or deprioritize others based on your business needs.


Step 4: Configure Your Product Finder Message Formats

The AI Product Finder presents results as a WhatsApp Carousel — a visually engaging format that displays product images, descriptions, and links. Customize the carousel to match your brand experience.

Available Configuration Options

Settings

Description

CTA Button Title

The call-to-action button label on each card. Maximum 25 characters.

Card Message

The message displayed on each carousel card. Use variables from your product catalogue to personalize it. This field cannot be left empty.

Number of Carousel Cards

The number of product cards the agent can send per response. Maximum 10 cards.

💡 Tip: Use product catalogue variables (e.g., product name, category, or brand) in your Card Message to make each recommendation feel dynamic and relevant.


Your AI Product Finder is now ready to help your customers discover the right products — conversationally, on WhatsApp.


🔗 Next Steps

  • Test your AI Product Finder in a sandbox conversation

  • Review your AI Persona settings to ensure tone consistency

  • Monitor agent performance in the Charles Analytics dashboard

Did this answer your question?