Alternative Solution for Scoring & Recommendation Services

Hi Meshkat Mohammadi,
Since Azure AI Personalizer is scheduled to be deprecated in 2026, you're right to look for alternative solutions for implementing scoring and recommendation functionality. One robust option is using Azure Machine Learning to build your own custom recommendation models. This gives you flexibility to apply techniques such as matrix factorization, content-based filtering, or hybrid approaches based on user context and historical behavior. Microsoft also provides tutorials and open-source tools like the Recommenders repository to help accelerate development.

Another option currently available—though also set to be retired in March 2026—is Azure Intelligent Recommendations, which provides automated and personalized recommendations based on user interactions and metadata. While it's still supported, it can be used to quickly integrate recommendation capabilities into your application. Additionally, Microsoft Fabric offers a unified analytics platform that can support end-to-end data processing and machine learning workflows. By leveraging Fabric's lakehouse architecture, you can efficiently manage and prepare your data for building and deploying recommendation systems. Together, these services offer flexible and scalable alternatives to meet your scoring and recommendation needs beyond Personalizer.
For more information: Tutorial: Create, evaluate, and score a recommendation system