# WooCommerce AI Product Recommendations on WhatsApp | ZazzyAgent

A customer rarely says:

> I want Product ID 1837.

They usually describe what they need.

For example:

> I need something for a birthday gift under ₹3,000.

An AI Agent can understand the request and use your connected WooCommerce store to find suitable products.

## What the AI Agent can do

A WooCommerce Sales Agent can:

*   Understand customer requirements
    
*   Ask useful follow-up questions
    
*   Search products
    
*   Recommend products
    
*   Explain why a product matches
    
*   Save customer preferences
    
*   Hand high-intent customers to sales staff
    

## What you need

Prepare:

*   Connected WhatsApp
    
*   Connected WooCommerce store
    
*   AI Agent
    
*   Product catalogue
    
*   WooCommerce product recommendation action
    

## Select the WooCommerce store

WooCommerce actions require the store to be configured first.

Use:

```text
##woocommerce_store## : 12
```

Select your connected store through the System Prompt autocomplete.

## Add product recommendations

Use:

```text
##woocommerce_recommend_products##
```

This allows the AI Agent to search and recommend products from the connected store.

## Ask useful questions

Don't force the customer through a long questionnaire.

Ask only what changes the recommendation.

For shoes:

*   Type of shoe
    
*   Intended use
    
*   Size
    
*   Budget
    

For electronics:

*   Purpose
    
*   Budget
    
*   Required features
    

## Example System Prompt

```text
You are the Sales Agent for our WooCommerce store.

Help customers find products that match their needs.

Ask short follow-up questions when important information
is missing.

When the customer asks for product recommendations,
search the connected WooCommerce store.

##woocommerce_recommend_products##

Recommend only products returned by WooCommerce.

Do not invent products, prices, discounts, availability,
features, or specifications.

Explain briefly why the recommended product matches
the customer's requirements.
```

## Example conversation

Customer:

> I need headphones for office use.

Agent:

> Sure. What's your approximate budget?

Customer:

> Around ₹5,000.

Agent:

> Do you prefer wireless headphones?

Customer:

> Yes.

Now the agent has enough information to search for relevant products.

## Don't recommend imaginary products

Your prompt should explicitly say:

> Recommend only products returned by the connected store.

This prevents the agent from creating products that don't exist.

## Don't invent prices

If the store returns:

> ₹4,999

the agent should use the actual information.

It should not say:

> I can give you 20% off.

unless that discount is actually available.

## Save product preference

You can combine product recommendations with customer data.

For example:

```text
When the customer clearly identifies a product category
they are interested in, save it to the Product Interest field.

##save_custom_field## : 42
```

See [How AI Agents Save Customer Information in ZazzyAgent](https://blog.zazzyagent.com/ai-agent-save-custom-fields-zazzyagent).

## Add a lead label

You could also classify a customer after strong purchase intent:

```text
When the customer clearly indicates strong purchase intent,
add the High Intent label.

##add_label## : 12
```

See [How to Add and Remove Customer Labels With ZazzyAgent AI Agents](https://blog.zazzyagent.com/ai-agent-add-remove-labels-zazzyagent).

## Hand high-intent customers to sales

A customer who says:

> I want to buy this one today.

may benefit from a human sales representative.

You can configure:

```text
When the customer clearly wants to purchase
and requires human sales assistance,
assign the conversation to the Sales team.

##assign_human_role## : 3
```

## Product knowledge vs WooCommerce search

These have different purposes.

**Knowledge**

> What does this product do?

**WooCommerce**

> Which products currently match this customer's requirements?

Static information can come from your knowledge source.

Live products should come from your connected ecommerce store.

## Common problems

### No products are returned

Check:

*   WooCommerce store is connected.
    
*   Correct store is selected.
    
*   Products exist.
    
*   The customer has provided enough information.
    

### Recommendations are irrelevant

Your prompt may need better instructions about what information should be collected before searching.

### AI invents products

Add:

> Only recommend products returned by WooCommerce.

### AI invents stock

Don't let the AI make live availability claims unless the connected store/action actually provides that information.

## A good recommendation workflow

```text
Customer describes need
        ↓
AI asks useful questions
        ↓
Product search
        ↓
WooCommerce results
        ↓
AI explains recommendations
        ↓
Customer chooses
        ↓
Sales / purchase workflow
```

The goal isn't to ask the customer 15 questions.

The goal is to understand enough to make a useful recommendation.