Boost your AI with Azure's new Phi model, streamlined RAG, and custom generative AI models

Microsoft announced significant updates to its Azure AI service, providing developers with more powerful and flexible options for building and deploying AI applications. These updates focus on three key areas: enhancements to Phi models, streamlined RAG, and custom generative AI models.

One of the most notable updates is the introduction of the new Phi-3.5-MoE model, a Mixture of Experts model that combines 16 smaller experts into one. This model delivers improvements in model quality and lower latency, providing users with the speed and computational efficiency of a small model with the domain knowledge and higher quality outputs of a larger model. Additionally, both the new MoE model and the Phi-3.5-mini model now support over 20 languages, allowing people to interact with the model in the language they are most comfortable using.

To streamline RAG pipelines for generative AI applications, Microsoft is offering integrated, end-to-end data preparation and embedding. With automatic vector indexing and querying using integrated access to embedding models, applications can unlock the full potential of the data available to them. Furthermore, custom generative extraction models in Azure AI Document Intelligence enable users to extract custom fields for unstructured documents with high accuracy, saving time and effort on manual annotation.

Finally, the general availability of Text to Speech (TTS) Avatar, a capability of the Azure AI Speech service, provides developers with a way to create engaging and personalized experiences for customers and employees. With TTS Avatar, businesses can create photorealistic avatars that appear natural with diverse voices, enhancing customer engagement and the overall experience.

In short, these updates to Azure AI represent a significant step forward in empowering developers to build more powerful and flexible AI applications. By providing a wide range of models and tools, Microsoft is making it easier for businesses to integrate AI into their operations and enhance customer experiences.