Implementing Synerise AI Search and AI recommendations

The Synerise AI Search and AI recommendation features are available from the Magento plugin version 3.3.2 and higher. We highly recommend to always update the plugin to the latest version.

You can use the Synerise AI Search engine to enhance product listings on category pages and improve the quick search functionality throughout your Magento store. Additionally, Synerise AI Recommendations can be integrated into various blocks on your website to deliver personalized product suggestions and boost user engagement.

This allows you to:

• Personalize search results
• Enable query suggestions and analyze their statistics
• Enable recommendations on a zero search stage
• Enable recommendations for no result search
• Insert Synerise AI recommendations in any place
• Analyze search statistics based on search events and analyze AI recommendations statistics based on recommendation events which will be available in a profile’s event history in Synerise

Prerequisites

• Integrate Synerise with Magento.

• Make sure your API key has the following permissions:

  • ITEMS_SEARCH_CONFIG_SEARCH_CREATE,
  • ITEMS_SEARCH_CONFIG_SEARCH_UPDATE,
  • ITEMS_SEARCH_CONFIG_SEARCH_READ,
  • ITEMS_SEARCH_SEARCH_READ,
  • ITEMS_SEARCH_RECENT_SEARCH_CREATE,
  • ITEMS_SEARCH_RECENT_SEARCH_READ,
  • API_MATERIALIZER_V2_RECOMMEND_CAMPAIGNS_RECOMMENDATIONS_READ

• If you want to enable query suggestions, create an index of suggestions in Synerise.

• In the Magento panel (Stores > Attributes > Product), review attributes which will be available for selection as searchable, filterable, and sortable attributes, and decide if you want to add some or remove them (after these actions always save the index configuration). You will always manage these attributes in Magento settings, and the plugin automatically transfers those settings to Synerise. During and after completing the integration, you should not change these settings in the Synerise search index used with your Magento shop.

  Option name	Value	Result
  Use in Search	Yes	An attribute is available on the list of searchable attributes
  Use in Layered Navigation	Yes	An attribute is available on the list of filterable (facetable) attributes
  Use in Search Results Layered Navigation	Yes	An attribute is available on the list of filterable (facetable) attributes
  Used for Sorting in Product Listing	Yes	An attribute is available on the list of sortable attributes

  

Procedure

The following steps must be performed in the given order:

1. Enable search item attribute synchronization (Magento)
2. Check if the catalog with Magento data exists in Synerise (Synerise)
3. Configure AI engine for the catalog (Synerise)
4. Prepare a search index (Magento)
5. Select attributes for the search index (Magento)
6. Select the search engine (Magento)

Optionally, you can:

• Define the settings for autocomplete search and query suggestions
• Inject Synerise recommendations into blocks
• Monitor performance of:
  • AI Search
  • AI Recommendations
  • Query suggestions

Enable search item attribute synchronization

In this part of the process, you will enable synchronization of item attributes used in search. These attributes will be available in Synerise in Data Modeling Hub > Catalogs, in a catalog named store-{id}, where {id} is the ID of the store.

1. In the Magento platform, go to Marketing > (Synerise) Configuration > Data.

2. Scroll to the Products section.

3. Recommended: Set the value of Include attributes used in search to Yes.
   This way all attributes used for searching (attributes selected as searchable, filterable, sortable in the configuration in Magento as a part of prerequisites) will be sent to Synerise catalog. We strongly recommend selecting this option when using Synerise AI Search.

4. Set Selectable attribute value to ID & Label.
   You can choose from two variants of this option:

   • Label & ID (array) - The attributes will be sent as an array.
   • Label & ID (attributes) - This setting allows you to send both the labels (for example, “red”, “shoes”, “XYZ brand”) and IDs of the attribute options (attribute_id). Selecting this option lets search users search and sort results using the labels.

   

5. To confirm the settings, click Save Config.

6. Clear configuration cache in System > Cache Management by performing the following steps:

   1. Select the checkbox next to Configuration cache type.
   2. Set the option above the table to Refresh.
   3. Confirm the action by clicking Submit.

   

7. Schedule full synchronization of data for the Product model to reflect the changes you made to the configuration.

Check if the catalog with Magento data exists in Synerise

The contents of the catalog will be synchronized with data from Magento. You can find it in Data Modeling Hub > Catalogs under the name in the following format: store-{id}, where {id} is the ID of the store. This catalog will be used in the next steps to create an item feed and configure AI engine for it.

1. In Synerise, go to Data Modeling Hub > Catalogs.
2. Check the list for the catalog named store-{id} (where {id} is the ID of the store).
3. If:
   • you find it, check whether its structure is consistent with the changes you made in “Enable search item attribute synchronization”. If not, wait until the catalog is updated.
   • the catalog isn’t created yet, wait until it appears on the list. The duration of creating a catalog depends on the data load.

Configure AI engine for the catalog

In this part of the process, you will create an item feed in Synerise based on the Synerise catalog and configure the AI engine for this feed (this lets you use AI Search and AI Recommendations).

1. To check if the item feed exists and the AI engine is configured, in Synerise, go to Settings > AI Engine Configuration.
   
   Important: Creating multiple feeds from the same catalogs is impossible.
2. Find the item feed on the list (its name will be in the following format: store-{id}). If it doesn’t exist, proceed to “I need to create an item feedn and configure the AI engine” section.
3. Enter the details of the item feed.
4. Proceed to “The item feed is already created”.

The item feed is already created

If you already created a feed from the store-{id} catalog, make sure that:

• the search engine is enabled for the existing feed in its configuration (see the Applied search engines section)
• recommendation models you want to use are enabled (only if you plan to use Synerise AI recommendations)
  Tip: You can find information about available recommendation models in Recommendation types. Recommendation types work based on the recommendation models with the same names as recommendation types.

I need to create an item feed and configure the AI engine

1. Go to Settings > Configuration > AI Engine Configuration.

2. In the upper right corner, click Add feed.
   Result: A pop-up appears.

3. On the pop-up, select Catalog.

4. As the catalog type, select Data catalog.

   

5. From the list of catalogs, select the catalog that receives synchronized data from Magento.
   The catalog’s name is store-{id}, where {id} is the ID of the store. If the catalog isn’t in the list, make sure that it’s not already used in another feed.

6. Confirm by clicking Apply.
   Result: The feed is created and added to the list.

7. On the list, find the feed you have just created and click the icon and then Edit.

8. On the Applied search engines section, click Show.

9. Enable the Search engines toggle.
   Result:

   

10. If you want to inject Synerise recommendations in your store:

    1. Enable the recommendation types you want to use.
    2. Wait for the recommendation model to finish training before creating recommendations. Instructions how to check the training status is available in Model status.
    3. Create recommendations and save their IDs, which you will need later.

11. Confirm the changes by clicking Apply.

Prepare search index in Magento

In this part of the process, you will create and/or select a Synerise search index for a store view. If you don’t have any indexes, creating one in the Magento plugin settings will also create it in Synerise, as described below.

1. Go to Marketing > (Synerise) Search Indexes.
2. Click Assign Search Index.
3. Select the store view for which the search will be applied.
4. To:
   • assign an existing index to the store view:
     1. Set Create to No.
     2. From Select, choose the index. The list contains the indexes from the Synerise platform ( AI Hub > Indexes)

        

   • create a new index:

     1. Set Create to Yes.

     2. In Name, enter the name of the index.

        

        Result: The index will also be available in the Synerise platform in AI Hub > Indexes.

Select attributes for the search index

In this part of the process, you will select searchable, filterable (facetable), and sortable attributes that will be available for use in search. This is the primary location for managing filterable, searchable, facetable, and sortable attributes. The scope of attributes available for selection is determined by the setup you defined as a part of the prerequisites.

1. Go to Marketing > (Synerise) Search Indexes.
2. Find the search index for which you want to select attributes.

   

3. For that index, from the Action column, click Setup.
4. Select searchable, filterable (facetable), and sortable attributes.

   

Select search engine

In this part of the process, you will select the Synerise search engine in the configuration of the Magento platform.

1. In Magento, go to Stores > Configuration > Catalog > Catalog Search.
2. In Search Engine, select Synerise AI Search.
3. Optionally, set Disable “Catalog Search” index to Yes.
   Such configuration removes unused search engine indexes from the reindex list.

   

Define the settings for autocomplete search

Autocomplete search suggests possible queries or products while users type, helping them find what they’re looking for faster. By predicting and displaying options in real-time, it streamlines the search process and improves user experience.

1. In the Magento administrator panel, go to Stores > Configuration.
2. Select the scope, for example Default Store View.
3. On the left panel, select Catalog > Catalog.
4. In the Catalog Search section, to define the length of the query after which search results show, in Minimal Query Length enter the number of characters of the query which triggers a search response.

   

5. Click Save Config.

Query suggestions

Suggestions are recommended options or predictions that appear while a user types a search query in the form of text strings. Displaying suggestions enhances the search experience by helping users find what they need faster. Enabling query suggestions will cause generation of the following events in the Synerise platform on the profile card of a user who uses search:

• suggestion.search event
• suggestion.search.click event

Before enabling this feature in Magento, you must create an index of suggestions in Synerise. After creating the index, it will be available for selection in the Magento panel.

1. In the Magento administrator panel, go to Marketing > Synerise > Configuration.
2. Select the scope, for example Default Store View.
3. On the left panel, select Synerise > AI Search.
4. In the Autocomplete section:
   1. To enable query suggestions, set Suggestion Enabled to Yes.
   2. To define the header of query suggestions in search results dropdown, in Suggestions Header, enter a value.
   3. Select the query suggestion index from Suggestions index
   4. To define the number of returned suggestion, in Suggestions Count, enter the number.

   

5. In the upper-left corner, click Save Config.
   Result:

   

Zero search recommendations

Zero search is the stage in the search journey when a user activates the search box but hasn’t typed a query yet. You can use this moment to promote products and enhance the search experience by:

• displaying product recommendations
• showing the user’s recent searches
• displaying top searches (most popular search queries)

To use product recommendations, you must first run a recommendation campaign in Synerise.

1. In the Magento administrator panel, go to Marketing > Synerise > Configuration.
2. Select the scope, for example Default Store View.
3. On the left panel, select Synerise > AI Search.
4. In the Autocomplete zero-state section:
   1. To enable recommendations and/or top searches, and/or recent searches when a user clicks the search box but before they type a query, set Zero-state enabled to Yes.
   2. To enable top searches, set Top searches enabled to Yes.
   3. To enable display of recently searched phrases, set Recent searches enabled, to Yes.
   4. To define the recommendation whose results will be displayed in the recommended products section in search box, in Campaign Id, enter the ID of the Synerise AI recommendation.
      To get the ID of the Synerise AI Recommendation, in Synerise, go to AI Hub > (AI Recommendations) Models. On the list, find the recommendation you want to use, click > Copy ID.
   5. To define the header of the section with recommended products, in Campaign title, enter the header.

      

5. In the upper-left corner, click Save Config.
   Result:

   

No results view

In the search journey, the no results view appears when a query returns no matches. Instead of showing an empty list, users can enable product recommendations, display top searches, and recently searched queries to keep visitors engaged and guide them to relevant content.

1. In the Magento administrator panel, go to Marketing > Synerise > Configuration.
2. Select the scope, for example Default Store View.
3. On the left panel, select Synerise > AI Search.
4. In the Autocomplete no-results section:
   1. To enable recommendations and/or top searches, and/or recent searches when a users query produces no results, set No-results enabled to Yes.
   2. To enable top searches, set Top searches enabled to Yes.
   3. To enable display of recently searched phrases, set Recent searches enabled, to Yes.
   4. To define the recommendation whose results will be displayed in the recommended products section in search box, in Campaign Id, enter the ID of the Synerise AI recommendation.
      To get the ID of the Synerise AI Recommendation, in Synerise, go to AI Hub > (AI Recommendations) Models. On the list, find the recommendation you want to use, click > Copy ID.
   5. To define the header of the section with recommended products, in Campaign title, enter the header.

      

5. In the upper-left corner, click Save Config.
   Result:

   

If you want to display AI recommendations on the no results page, you can inject Synerise recommendations into blocks which you can use on the no results page.

Injecting Synerise recommendations into blocks

You can display Synerise AI recommendations at any place in your store. In this part of the process, you will create a block in the Magento admin panel which contains a Synerise product recommendation widget type. Then, you will insert this block in an existing page configuration.

1. In the Magento admin panel, go to Content > Blocks > Add New Block.
   Result:

   

2. Fill out the following fields: Block Title and Identifier.

3. Select the store view for which the block will be available.

4. In the bottom-right corner, click Edit with Page Builder.
   Result:

   

5. On the left panel, expand Elements and select HTML Code.

6. Hover the cursor over the HTML code widget you added.

7. Click HTML Code.
   Result:

   

8. Click Insert Widget.

9. From the Widget Type dropdown list, select Synerise Product Recommendations.
   Result: Widget Options configuration appears.

   

10. In the Campaign index field, enter the ID of the recommendation.

11. Confirm by clicking Insert widget in the upper-right corner.

12. Confirm by clicking Save in upper-right corner.

13. Confirm by clicking Save as Template.

14. Confirm by clicking Save in upper-right corner.

15. Go to Content > Pages.

16. Find the page where you want to inject recommendation, go to its editing mode.

    

17. Expand the Content section.
    Result:

    

18. Click Edit with Page Builder.

19. On the left panel, go to Add Content > Block.

20. Drag and drop a block on the canvas.

21. Go to the settings of the block and select the block with the recommendation you created.

22. Save the changes.

Defining advanced AI search settings in the Synerise platform

This part of the process is optional.

To define advanced options which let you optimize search engine performance:

1. In Synerise, go to AI Hub > Indexes.
2. Find the search index you selected in “Prepare a search index in Magento”.
3. Open the details of the search index by clicking its name on the list.
4. You can edit or define the following options:

• Item ranking for sorting search results

• Management of typos and missing words

• Response, filterable, facetable, and sortable attributes
  

  WARNING: The primary location for managing filterable, searchable, facetable, and sortable attributes is the Magento panel. While deleting these attributes in Synerise will take immediate effect, any additions of new attributes must be done through the Magento panel. To ensure compatibility, it is important to maintain the same scope of attributes on both platforms when managing them in the Magento panel and Synerise.

  The following attributes are required:

  • Response attributes:

    • entity_id

  • Searchable attributes:

    • name

  • Filterable attributes

    • entity_id
    • is_salable
    • visibility
    • deleted

  • Facetable attributes

    • category_ids
    • entity_id
    • is_salable
    • visibility
    • deleted

• Rules

• Search language

• Search method, searchable attributes, and search relevance

• Synonyms

• A/B testing

Monitoring Synerise AI Search performance

You can monitor the performance of AI Search, you can find out about the metrics and how to get to them in AI Search Engine statistics.

Monitoring query suggestions performance

If you enabled query suggestions for the search, you can monitor their statistics. You can find out about the metrics and how to get to them in Query suggestion statistics.

Monitoring Synerise AI Recommendations performance

You can monitor the performance of the AI recommendations used on your store, you can find out about the metrics and how to get to them in Recommendation statistics.