AI module improvements

Today we wanted to highlight a few key improvements introduced to the AI module that bring additional value and expand the capabilities of individual functionalities.

Let's take a closer look at each of these enhancements.

Disabling typo tolerance

Typo tolerance is a feature that helps search engines handle small mistakes in user queries, like misspelled words. However, there are certain cases where there should be a possibility to turn it off.

For example, when searching for an item using its SKU, it is important to find an exact match. If the search engine automatically applies a typo tolerance and there is no exact match in the index, it may return irrelevant results. Similar cases arise with numeric attributes, such as postal codes, prices, and sizes, where exact matches are preferred.

No alt text provided for this image

To address these needs, the improvement introduced to the AI Search allows you to specify which searchable attributes/words should be exact matches, disabling typo tolerance for those attributes/words. This means that if a search is performed on an attribute/word specified as an exact match, the search engine will only return results that precisely match the input, without applying any typo tolerance.

In addition, there is an option to globally disable typo tolerance for all numeric attributes. This ensures that numeric attributes, such as SKU and postal codes, for example, are not subject to typo tolerance unless explicitly specified otherwise.

These enhancements provide greater control over when typo tolerance is applied, enabling fine-tuning of search results based on specific attribute requirements, keywords and ensuring accurate matches for numeric values.

For more details on handling typos, visit our User Guide.  

Multiple predicted events in Time Optimizer

No alt text provided for this image

The enhancement introduced to Time Optimizer allows it to consider a wide range of predicted events. This means that Time Optimizer can now take into account a wider range of anticipated activities and behaviors when determining the optimal time to reach customers.

By incorporating multiple predicted events, Time Optimizer gains a deeper understanding of customer engagement across various channels. It can analyse patterns and preferences more comprehensively, allowing it to choose the ideal time to send emails or perform other customer-oriented activities.