Guides / Personalization / AI Personalization (beta) / What is AI Personalization?

How does AI Personalization work?

AI Personalization is a beta feature according to Algolia’s Terms of Service (“Beta Services”).

AI Personalization builds user profiles for each of your users based on their behavior on your website or app. A user’s behavior implicitly reveals their preferences and interests and can be referred to as “affinities”.

Identifying categories as facets

Most users tend to have affinities for categories of items rather than specific individual items. For example, a user can have an affinity for athletic wear while shopping for clothing, fictional works when browsing through books, or pop music when listening to music. These categories make up the facets within your data. To build accurate user profiles for your users, AI Personalization needs to understand which categories, that is, facets, your user has an affinity for.

Take a music streaming service as an example. For the streaming service to make search results more personalized to each user, it’s more important for it to understand the music categories (genres and artists) the user likes, rather than the specific song titles they’ve listened to. AI Personalization works by creating a user profile that details musical preferences (affinities) based on the categories of music they like. This profile can be used in conjunction with Algolia’s NeuralSearch or keyword search to personalize the user’s search results.

Translating user behavior into affinities through events

User behavior describes all search and non-search interactions users have with your website or app. Meaningful user behavior can provide insights into a user’s affinities. Examples of meaningful behavior include when a user clicks, views, likes, favorites, bookmarks, or purchases a particular item or category.

User behavior reaches AI Personalization in the form of standardized events. Each time a user engages in meaningful behavior, you need to send an event to Algolia so that AI Personalization can use it to build that user’s profile. It’s important to gather as much data as possible when trying to build the most comprehensive and relevant user profiles for your users.

Understanding user profiles

A user profile is a representation of a user based on their behavior on your website or app. It is the main entity that is used in conjunction with Algolia’s NeuralSearch or Keyword Search to personalize search results. It’s defined by a JSON format containing an identifier for your user and their affinities.

Sending events accurately is the first step to building relevant user profiles. Algolia requires events that indicate how users are interacting with your app. For example, an online bookstore could send events that map to a user clicking on a specific book within a listing, viewing information about the genre and author of the book on a detail page, adding the book to their cart, and purchasing the book.

Once events have been implemented and are being sent, AI Personalization applies different algorithms to your data (index and events). Patterns and relationships are identified within your data to predict user affinities. These affinities are used to create a user profile, that’s continually updated with new data. This helps AI Personalization adapt to changes in user behavior and preferences.

Did you find this page helpful?