Announcing updates to Azure AI Search to help organizations build and scale generative AI applications

6 months ago 28
News Banner

Looking for an Interim or Fractional CTO to support your business?

Read more

Today we are announcing important changes to Azure AI Search successful enactment for customers gathering accumulation acceptable generative AI applications. Azure AI Search has drastically accrued retention capableness and vector scale size astatine nary further cost, truthful customers tin tally retrieval augmented procreation (RAG) astatine immoderate scale, without having to compromise outgo oregon performance. 

In this post, we volition locomotion done however customers:

  • Can execute much scalability astatine a little outgo with today’s changes.
  • Trust Azure AI Search to grip their ample RAG workloads. 
  • Apply precocious hunt strategies to navigate analyzable information to innovate successful ways antecedently unimaginable. 
a idiosyncratic   sitting astatine  a table  utilizing a computer

Azure AI Search

Deliver accurate, hyper-personalized responses successful your generative AI applications

Announcing much scalability and show astatine a little outgo with Azure AI Search

Azure AI Search has importantly raised vector and retention capacity, offering customers greater scalability, precocious performance, and much information per dollar.

The magnitude of capableness and compute disposable has accrued for Azure AI Search’s Basic and Standard tiers, successful select regions.

Here’s what this means.

Users volition now spot up to a:

  • 11x summation successful vector scale size
  • 6x summation successful full storage.
  • 2x betterment successful indexing and query throughput

With these changes, customers tin present precocious prime experiences for each idiosyncratic and interaction, astatine immoderate scale. Customers tin standard their generative AI applications to a multi-billion vector scale successful a azygous hunt instance, without compromising velocity and performance.

Supporting ample RAG-based applications with a trusted endeavor retrieval system 

Over fractional of Fortune 500 companies spot Azure AI Search to negociate their ngo captious endeavor hunt and generative AI applications. OpenAI, Otto Group, KPMG, and PETRONAS usage Azure AI Search to enactment retrieval augmented procreation (RAG) workloads. 

When OpenAI announced their RAG-powered “GPTs” and the Assistant API astatine OpenAI DevDay 2023, OpenAI needed to guarantee their retrieval strategy was susceptible of handling unprecedented request and scale. OpenAI turned to Azure AI Search for its capableness to enactment their large, internet-scale RAG workloads. 

Azure AI Search present supports RAG capabilities for ChatGPT, GPTs, and the Assistant API, and provides hunt functionality to products similar the GPT Store. Any clip idiosyncratic searches successful oregon adds a record to 1 of these products, Azure AI Search is the retrieval strategy that makes it happen. 

ChatGPT alone, as of November 2023, receives 100 cardinal play visitors, with implicit 2 million developers gathering with its API. Within 2 months of announcing customized GPTs, 3 million GPTs were created. These are monolithic numbers, with users spanning crossed the globe. Truly RAG astatine scale.

Building amended applications with a modern, precocious retrieval system

Teams successful nonrecreational services, healthcare, and telecommunications person recognized that successful bid to physique a generative AI exertion that performs arsenic its designed, using lone 1 hunt practice similar vector search, simply doesn’t work.

Different retrieval techniques execute amended for antithetic usage cases. High prime retrieval systems harvester aggregate techniques to screen the assortment of scenarios that immoderate fixed exertion tends to see. 

Azure AI Search tin alteration applications to use a assortment of approaches close retired of the box, including hybrid retrieval and semantic reranking, enabling developers to execute objectives faster and much effectively. 

Telus Health uses precocious RAG to present a lawsuit attraction application 

Telus Health, headquartered successful Canada, is astatine the forefront of providing technology-driven solutions and services to employers, individuals, healthcare practitioners, and insurers. The enactment introduced a lawsuit enactment platform, designed to reply idiosyncratic inquiries astir circumstantial wellness plans, and connection guidance for navigating their website. The archetypal implementation, which relied solely connected vector search, fell abbreviated successful addressing each requirements wrong a unified system. Because of this, Telus Health turned to Azure AI Search, known for its innovative, broad acceptable of hunt technologies.

The Guide Team astatine Telus Health was instrumental successful designing their hunt strategy and leveraging AI Search efficaciously to heighten the platform. By expanding their retrieval strategy, and implementing hybrid hunt with semantic reranking, Telus Health enabled the strategy to efficiently grip some questions astir lawsuit circumstantial documents and utilizing the organization’s website. This strategical enhancement, supported by Azure AI Search, has importantly improved the platform’s accuracy and responsiveness, and showcases Telus Health’s committedness to delivering exceptional lawsuit support. 

NIQ Brandbank empowers brands to optimize their online beingness with multi-vector retrieval 

NIQ Brandbank supports Fast-Moving Consumer Goods (FMCG) brands with solutions to present rich, applicable contented and imagery for their integer shelf, successful bid to outshine their competitors.

NIQ Brandbank’s solution Content Health+ empowers brands to optimize their online beingness with data-driven, actionable guidance and accusation to amusement however their merchandise contented stacks up against marketplace rivals. 

With their simple, user-friendly format, the exertion helps brands amended merchandise placement crossed retailer hunt results, boost income and elevate their online presence. 

Content Health+ pulls from probe completed by a NIQ Data Impact team, to place what merchandise attributes power integrated placement connected the integer shelf. On the backend, the exertion searches this probe stored successful some images and text, utilizing multi-vector search, and surfaces the astir applicable results with hunt reranking. This functionality provides prime recommendations for what contented attributes a marque should absorption connected to amended their show and bottommost line.

For this exertion to execute arsenic designed, Content Health+ was built utilizing hybrid multi-vector hunt and semantic ranking. For ecommerce and proposal applications, much ideas and opportunities are realized by combining assorted retrieval methods. 

Azure AI Search allows america to usage hybrid multi-vector search, utilizing substance and representation embeddings with semantic ranking to beforehand the astir semantically applicable products to the top. From accumulation acceptable automatic information ingestion from Azure information sources to integration with Azure Machine Learning, Azure AI Search is precisely what we needed to marque the Content Health+ exertion a reality.” 

—Gabriel Harris PhD, Principal Data Scientist 

Learn much astir Azure AI Search

With these announcements today, we are making it easier for the AI systems to retrieve accusation astatine scale. Customers tin innovate with assurance with state-of-the-art retrieval exertion successful Azure AI Search and an enterprise-ready foundation.  

For much connected RAG and Azure AI Search: 

The station Announcing updates to Azure AI Search to assistance organizations physique and standard generative AI applications appeared archetypal connected Microsoft Azure Blog.

Read Entire Article