3 Questions About Recommendations AI

[Batucan]

Hello everyone, i have a few question about GCP Recommendations AI?

Let's say we train a model using Recommendations AI's User Interface. What I'm wondering is, when a user buys a product, does the system automatically stop recommending that product for that customer or does it continue to recommend the same product?

Secondly, is it possible to create configs and controls via user event attributes? Is it possible for certain rules to affect the recommendation when a user triggers an event (e.g. purchase, add to cart, etc.)?

Finally, does a model developed with Recommendations AI use any cookies in the entire end-to-end process, from importing the data to deploying the model and taking it live? If so, is it possible to perform this action without using any cookies?

Thanks to your help.

[Poala_Tenorio]

In Recommendations AI, when a user buys a product, the system typically adjusts its recommendations based on user behavior. While it's not guaranteed to immediately stop recommending the purchased product, the system often considers the user's preferences and might lower the priority of that specific item in subsequent recommendations. However, the exact behavior can be customized and fine-tuned based on your configuration.

Regarding your second question, Recommendations AI offers some flexibility in configuring recommendations based on user events. You can set up rules and controls that influence recommendations when certain events occur, like a purchase or adding an item to the cart. These event-based rules can impact the recommendations presented to the user by adjusting their preferences or altering the recommendation logic for subsequent interactions.

As for the use of cookies, Recommendations AI doesn't explicitly rely on cookies in its end-to-end process. The system primarily operates by analyzing user behavior and interaction data to generate recommendations. However, if your implementation involves tracking user behavior through cookies to gather data for the model, that would be a separate aspect that you'd need to handle within your application or website. It might be possible to use Recommendations AI without employing cookies for tracking user behavior if the necessary data can be collected through other means or if your use case doesn't rely on individual user tracking.

Keep in mind that while Recommendations AI itself might not heavily depend on cookies, the implementation within your broader application or platform might involve cookie usage for data collection, user identification, or session management, depending on your specific setup and requirements.