New AI capabilities available to Algolia customers — Query Categorization

Query Categorization Launches in Public Beta

Query Categorization is now available in public beta as of September 27, 2022. Algolia customers under the premium plan can access Query Categorization directly from their Algolia dashboard and experience this new and impactful capability for themselves. We view Query Categorization as a critical building block of our Search and Discovery Platform. This new capability will dramatically reduce end-user effort by making it easier for your customers to search, discover, and find what they are looking for and more.

With Algolia Query Categorization, developers, merchandisers, media companies, and content providers can anticipate their customer’s intent. Connect your customer to the most relevant category or selection of items, and help them design the content or product categories that best meet their needs throughout every moment of the end-user journey. Product Managers and Merchandisers alike can leverage Algolia to open the doors to targeted relevance and improved website conversion rates. Query Categorization offers the following capability keys:

• Automatic Filtering & Boosting – Leverage category predictions to automatically filter or boost search records. This feature reduces time to conversion and improves your customer’s experience with increased search precision.
• Analytics Grouped by Predicted Category – Quickly identify opportunities by comparing queries to search results. Algolia clusters queries and predicts categories to compare. These insights help uncover underperforming searches.
• Prediction access through the Search API – With this new AI model, associate each query with a category from the catalog queries, e.g., map the query “banana” to “fresh produce” > “fruits.” Customers can tailor their search around the predicted category to drive more precise results.
• Rules (coming soon!) – Rules relieve the manual relevance tweaking by creating rules based on the query predicted category. Rules deliver a unified experience across all queries from the same category; for example, jeans, trousers, and pants can have the same shopping experience applied simultaneously

My experience with managing product categories comes from working in Corporate Finance in the CPG space. I spent a tremendous amount of time in databases, analytics tools, and spreadsheets grouping product SKUs to aggregate, or roll up into classified product categories to allocate costs, classify revenue, and improve margins with promotions and product placement. Your category manager is doing much of the same without a team of data scientists and AI experts backing them up. Algolia’s AI model allows you to automate manual processes, unlocks automatic boosting & burying, delivers higher-level analytics, and promotes more efficient merchandising.

I have seen businesses attempt to replicate the behemoths of the online retail world by building highly manual processes to deliver search results that match them with ranked products and categories. Now comes the great equalizer – Algolia Query Categorization. With technical parity, you can compete against the leading retailers with just a few clicks.

Algolia’s API-first platform delivers powerful query extraction and classification techniques. It empowers developers to quickly build and deploy delightful end-user experiences with the predicted results. You can dynamically map a customer’s search to an aligned category at the enterprise scale. Algolia’s Query Categorization AI model is designed to display customer intent. Algolia focuses on helping business users to finetune site and search results while increasing cart sizes, optimizing conversion rates, and improving customer satisfaction.

How?

Send click and conversion data to Algolia, specify how you want to classify your records in the Query Categorization section of the Algolia dashboard, and let the model automatically learn how to categorize the queries that your end-users are making. You will be able to access predictions for each search query based on which categories it is most likely associated with, as a consequence, you will be able to  improve your search experience with a set of no-code tools and features.

Are you looking for search trends and behaviors? 

Query Categorization delivers search trends and behaviors by automatically predicting product categories based on previous searches performed by your customers. As a vegetarian looking for plant-based products, I visit my favorite health food grocer’s website and type in “burgers” in the search bar. Algolia determines the likelihood that I am searching for options in the “Products / Meat / Meat Alternatives” category rather than the “Products / Meat / Beef” category. How did Algolia know that? Because my favorite grocery store serves a huge majority of vegans and vegetarians. Algolia helps the grocer clearly understand their customer’s preferences and boost results accordingly.

To achieve this intent anticipation, Query Categorization leverages a vector-based semantic model to predict which classifications a search query is most likely to align with. This (Our) probabilistic model provides developers, content, and product managers the power to anticipate what their customers are looking for during their search. The result is higher customer satisfaction, stronger brand loyalty and as result  more sales making you a ‘category leader’ and making your Finance team notice your contribution to the business bottom line.

---

For more information, check out the educational content that we’ve put together, as well as our Query Categorization documentation.