Microsoft has revealed a huge increase in storage capacity and vector index size for its Azure AI Search platform that comes at no additional price to customers using the AI-augmented search platform.
Users will benefit from an 11-fold increase in vector index, as well as a six-fold increase in total storage and a two-fold improvement in both indexing and query performance, the tech giant confirmed.
Microsoft said the move aims to provide developers with expanded capabilities to scale generative AI applications by leveraging “multi-billion dollar vector search.”
The upgrade also allows enterprises to run Recovery Augmented Generation (RAG) at scale without needing to compromise on price, using “more data per dollar.”
Azure AI Search now supports RAG capabilities for ChatGPT, GPT, and Assistant API, acting as the recovery system that ensures the continued functionality of the GPT store.
Microsoft said the size of ChatGPT's user base indicates the possibilities of Azure AI Search when it comes to RAG at scale, and other large-scale companies such as Otto Group and KPMG have also reported using the platform.
RAG systems are a necessity in the development of generative AI applications
These more advanced instances of RAG systems can be applied in numerous places across the global economy, Microsoft said, with professional services, healthcare and telecommunications teams citing the importance of vector search in generative AI.
These groups have found that to create a generative AI application that works as designed, “using only one search practice like vector search simply doesn't work.”
Higher quality recovery systems cover a greater variety of scenarios for application functionalities, the firm noted.
These approaches include hybrid retrieval and semantic reclassification, features that help developers achieve their goals more effectively and efficiently.
Microsoft took note of two companies that are currently showcasing the possibilities created by the pursuit of Azure AI: Telus Health, which uses advanced RAG to offer a customer service application, and NIQ Brandbank, which is optimizing its online presence through recovery of multiple vectors.
The Telus Health team was able to leverage AI search to improve the platform, expand its retrieval strategy, and implement hybrid search and semantic reclassification to improve question handling.
NIQ Brandbank takes a different approach to RAG as an enterprise solution, using multi-vector search to analyze research on the backend for use in its Content Health+ platform.
