News and advancements from Microsoft Ignite to showcase our committedness to your occurrence successful this dynamic era. Let’s get started.

In the midst of this unthinkable technological shift, 2 things are clear: organizations are seeing tangible results from AI and the innovation imaginable is limitless. We purpose to empower YOU—whether arsenic a developer, IT pro, AI engineer, concern determination maker, oregon a information professional—to harness the afloat imaginable of AI to beforehand your concern priorities. Microsoft’s endeavor experience, robust capabilities, and steadfast commitments to trustworthy exertion each travel unneurotic successful Azure to assistance you find occurrence with your AI ambitions arsenic you make the future. 

This week we’re announcing quality and advancements to showcase our committedness to your occurrence successful this dynamic era. Let’s get started.

Introducing Microsoft Azure AI Foundry: A unified level to design, customize, and negociate AI solutions 

Every caller procreation of applications brings with it a changing acceptable of needs, and conscionable arsenic web, mobile, and unreality technologies person driven the emergence of caller exertion platforms, AI is changing however we build, run, govern, and optimize applications. According to a Deloitte report, astir 70% of organizations person moved 30% oregon less of their Generative AI experiments into production—so determination is simply a batch of innovation and results acceptable to beryllium unlocked. Business leaders are looking to trim the clip and outgo of bringing their AI solutions to marketplace portion continuing to monitor, measure, and measure their show and ROI.

This is wherefore we’re excited to unveil Azure AI Foundry contiguous arsenic a unified exertion level for your full enactment successful the property of AI. Azure AI Foundry helps span the spread betwixt cutting-edge AI technologies and applicable concern applications, empowering organizations to harness the afloat imaginable of AI efficiently and effectively.

We’re unifying the AI toolchain in a caller Azure AI Foundry SDK that makes Azure AI capabilities accessible from acquainted tools, similar GitHub, Visual Studio, and Copilot Studio. We’ll besides germinate Azure AI Studio into an enterprise-grade absorption console and portal for Azure AI Foundry.

Azure AI Foundry is designed to empower your full organization—developers, AI engineers, and IT professionals—to customize, host, run, and negociate AI solutions with greater easiness and confidence. This unified attack simplifies the improvement and absorption process, helping each stakeholders absorption connected driving innovation and achieving strategical goals.

For developers, Azure AI Foundry delivers a streamlined process to swiftly accommodate the latest AI advancements and absorption connected delivering impactful applications. Developers volition besides find an enhanced experience, with entree to each existing Azure AI Services, and tooling on with caller capabilities we’re announcing today. 

For IT professionals and concern leaders, adopting AI technologies raises important questions astir measurability, ROI, and ongoing optimization. There’s a pressing request for tools that supply wide insights into AI initiatives and their interaction connected the business. Azure AI Foundry enables leaders to measurement their effectiveness, align them with organizational goals, and much confidently put successful AI technologies.

To assistance you standard AI adoption successful your organization, we’re introducing broad guidance for AI adoption and architecture wrong Azure Essentials so you are equipped to successfully navigate the gait of AI innovation. Azure Essentials gives you entree to Microsoft’s champion practices, merchandise experiences, notation architectures, skilling, and resources into a azygous destination. It’s a large mode to payment from each we’ve learned and the attack you’ll find aligns straight with however to marque the astir of Azure AI Foundry.

In a marketplace flooded with disparate technologies and choices, we created Azure AI Foundry to thoughtfully code divers needs crossed an enactment successful the pursuit of AI transformation. It’s not conscionable astir providing precocious tools, though we person those, too. It’s astir fostering collaboration and alignment betwixt method teams and concern strategy.

Now, let’s dive into further updates designed to heighten the wide acquisition and ratio passim the AI improvement process, nary substance your role.

Introducing Azure AI Agent Service to automate concern processes and assistance you absorption connected your astir strategical work  

AI agents person immense imaginable to autonomously execute regular tasks, boosting productivity and efficiency, each portion keeping you astatine the center. We’re introducing Azure AI Agent Service to assistance developers orchestrate, deploy, and standard endeavor AI-powered apps to automate concern processes. These intelligent agents grip tasks independently, involving quality users for last reappraisal oregon action, ensuring your squad tin absorption connected your astir strategical initiatives. 

A standout diagnostic of Agent Service is the quality to easy link endeavor information for grounding, including Microsoft SharePoint and Microsoft Fabric, and tools integration to automate actions. With features similar bring your ain retention (BYOS) and backstage networking, it ensures information privateness and compliance, helping organizations support their delicate data. This allows your concern to leverage existing information and systems to make almighty and unafraid agentic workflows.

Enhanced observability and collaboration with a caller absorption halfway experience

To enactment the improvement and governance of generative AI apps and fine-tuned models, contiguous we’re unveiling a caller absorption halfway acquisition close successful Azure AI Foundry portal. This diagnostic brings indispensable subscription information, specified arsenic connected resources, entree privileges, and quota usage, into 1 pane of glass. This tin prevention improvement teams invaluable clip and facilitate easier information and compliance workflows passim the full AI lifecycle. 

Expanding our AI exemplary catalog with much specialized solutions and customization options 

From generating realistic images to crafting human-like text, AI models person immense potential, but to genuinely harness their power, you request customized solutions. Our AI exemplary catalog is designed to supply prime and flexibility and guarantee your enactment and developers person what they request to research what AI models tin bash to beforehand your concern priorities. Along with the latest from OpenAI and Microsoft’s Phi household of tiny connection models, our exemplary catalog includes unfastened and frontier models. We connection much than 1,800 options and we’re expanding to connection adjacent much tailored and specialized task and industry-specific models.  

We’re announcing additions that see models from Bria, present successful preview, and NTT DATA, present mostly available. Industry-specific models from Bayer, Sight Machine, Rockwell Automation, Saifr/Fidelity Labs, and Paige.ai are besides disposable today successful preview for specialized solutions successful healthcare, manufacturing, finance, and more.

We’ve seen Azure OpenAI Service depletion much than treble implicit the past six months, making it wide customers are excited astir this concern and what it offers_². We look guardant to bringing much innovation to you with our partners astatine OpenAI, starting with new fine-tuning capabilities similar imaginativeness fine-tuning and distillation workflows which let a smaller exemplary similar GPT-4o mini to replicate the behaviour of a larger exemplary specified arsenic GPT-4o with fine-tuning, capturing its indispensable cognition and bringing caller efficiencies.

Along with unparalleled exemplary choice, we equip you with indispensable tools similar benchmarking, evaluation, and a unified exemplary inference API truthful you tin explore, compare, and prime the champion exemplary for your needs without changing a enactment of code. This means you tin easy swap retired models without the request to recode arsenic caller advancements emerge, ensuring you’re ne'er locked into a azygous model.

New collaborations to streamline exemplary customization process for much tailored AI solutions

We’re announcing collaborations with Weights & Biases, Gretel, Scale AI, and Statsig to accelerate end-to-end AI exemplary customization. These collaborations screen everything from information mentation and procreation to training, evaluation, and experimentation with fine-tuned models. 

The integration of Weights & Biases with Azure volition supply a broad suite of tools for tracking, evaluating, and optimizing a wide scope of models successful Azure OpenAI Service, including GPT-4, GPT-4o, and GPT-4o-mini. This ensures organizations tin physique AI applications that are not lone powerful, but besides specifically tailored to their concern needs.  

The collaborations with Gretel and Scale AI purpose to assistance developers region information bottlenecks and marque information AI-ready for training. With Gretel Azure OpenAI Service integration, you tin upload Gretel generated information to Azure OpenAI Service to fine-tune AI models and execute amended show successful domain-specific usage cases. Our Scale AI concern volition besides assistance developers with adept feedback, information preparation, and enactment for fine-tuning and grooming models. 

The Statsig collaboration enables you to dynamically configure AI applications and tally almighty experiments to optimize your models and applications successful production. 

More RAG show with Azure AI Search 

Retrieval-augmented generation, oregon RAG, is important for ensuring accurate, contextual responses and reliable information. Azure AI Search present features a generative query motor built for precocious show (for prime regions). Query rewriting, disposable successful preview, transforms and creates aggregate variations of a query utilizing an SLM-trained (Small Language Model) connected information typically seen successful generative AI applications. In addition, semantic ranker has a caller reranking model, trained with insights gathered from lawsuit feedback and manufacture marketplace trends from implicit a year.  

With these improvements, we’ve shattered our ain show records—our caller query motor delivers up to 12.5% amended relevance, and is up to 2.3 times faster than past year’s stack. Customers tin already instrumentality vantage of amended RAG show today, without having to configure oregon customize immoderate settings. That means improved RAG show is delivered retired of the box, with each the hard enactment done for you.

Effortless RAG with GitHub models and Azure AI Search—just adhd data 

Azure AI Search volition soon powerfulness RAG successful GitHub Models, offering you the aforesaid casual entree glide way to bring RAG to your developer situation successful GitHub Codespaces. In conscionable a fewer clicks, you tin experimentation with RAG and your data. Directly from the playground, simply upload your information (just resistance and drop), and a escaped Azure AI Search scale volition automatically beryllium provisioned. 

Once you’re acceptable to build, copy/paste a codification snippet into your dev situation to adhd much information oregon effort retired much precocious retrieval methods offered by Azure AI Search. 

This means you tin unlock a full-featured cognition retrieval strategy for free, without ever leaving your code. Just adhd data.

Advanced vector hunt and RAG capabilities present integrated into Azure Databases  

Vector hunt and RAG are transforming AI exertion improvement by enabling much intelligent, context-aware systems. Azure Databases present integrates innovations from Microsoft Research—DiskANN and GraphRAG—to supply cost-effective, scalable solutions for these technologies.

GraphRAG, disposable successful preview successful Azure Database for PostgreSQL, offers precocious RAG capabilities, enhancing ample connection models (LLMs) with your backstage PostgreSQL datasets. These integrations assistance empower developers, IT pros, and AI engineers alike, to physique the adjacent procreation of AI applications efficiently and astatine unreality scale. 

DiskANN, a state-of-the-art suite of algorithms for low-latency, highly scalable vector search, is present generally disposable successful Azure Cosmos DB and successful preview for Azure Database for PostgreSQL. It’s besides combined with full-text hunt to powerfulness Azure Cosmos DB hybrid search, presently successful preview.  

Equipping you with liable AI tooling to assistance guarantee information and compliance  

We proceed to backmost up our Trustworthy AI commitments with tools you tin use, and contiguous we’re announcing 2 more: AI reports and risk and information evaluations for images. These updates assistance guarantee your AI applications are not lone innovative, but harmless and compliant. AI reports alteration developers to papers and stock the usage case, exemplary card, and valuation results for fine-tuned models and generative AI applications. Compliance teams tin easy review, export, approve, and audit these reports crossed their organization, streamlining AI plus tracking, and governance. 

We are besides excited to denote caller collaborations with Credo AI and Saidot to enactment customers’ end-to-end AI governance. Credo AI pioneered a liable AI level enabling broad AI governance, oversight, and accountability. Saidot’s AI Governance Platform helps enterprises and governments negociate hazard and compliance of their AI-powered systems with ratio and precocious quality. By integrating the champion of Azure AI with innovative AI governance solutions, we anticipation to supply our customers with prime and foster greater cross-functional collaboration to align AI solutions with their ain principles and regulatory requirements.   

Transform unstructured information into multimodal app experiences with Azure AI Content Understanding  

AI capabilities are rapidly advancing and expanding beyond accepted substance to amended bespeak contented and input that matches our existent world. We’re introducing Azure AI Content Understanding to marque it faster, easier, and much cost-effective to physique multimodal applications with text, audio, images, and video. Now successful preview, this work uses generative AI to extract accusation into customizable structured outputs.  

Pre-built templates connection a streamlined workflow and opportunities to customize outputs for a wide scope of use-cases—call halfway analytics, selling automation, contented search, and more. And, by processing information from aggregate modalities astatine the aforesaid time, this work tin assistance developers trim the complexities of gathering AI applications portion keeping information and accuracy astatine the center.

Advancing the developer acquisition with caller AI capabilities and a idiosyncratic usher to Azure 

As a institution of developers, we ever support the developer assemblage apical of caput with each advancement we bring to Azure. We strive to connection you the latest tech and champion practices that boost impact, acceptable the mode you work, and amended the improvement acquisition arsenic you physique AI apps. 

We’re introducing 2 offerings successful Azure Container Apps to assistance alteration however AI app developers work: serverless GPUs, present successful preview, and dynamic sessions, disposable now.  

With Azure Container Apps serverless GPUs—you tin seamlessly tally your lawsuit AI models connected NVIDIA GPUs. This diagnostic provides serverless scaling with optimized acold start, per-second billing, with built-in standard down to zero erstwhile not successful use, and reduced operational overhead. It supports casual real-time inferencing for customized AI models, allowing you to absorption connected your halfway AI codification without worrying astir managing GPU infrastructure. 

Azure Container Apps dynamic sessions—offer accelerated entree to unafraid sandboxed environments. These sessions are cleanable for moving codification that requires beardown isolation, specified arsenic ample connection exemplary (LLM) generated codification oregon extending and customizing bundle arsenic a work (SaaS) apps. You tin mitigate risks, leverage serverless scale, and trim operational overhead successful a cost-efficient manner. Dynamic sessions travel with a Python codification interpreter pre-installed with fashionable libraries, making it casual to execute communal codification scenarios without managing infrastructure oregon containers. 

These caller offerings are portion of our ongoing enactment to enactment Azure’s broad dev capabilities wrong casual reach. They travel close connected the heels of announcing the preview of GitHub Copilot for Azure, which is similar having a idiosyncratic usher to Azure. By integrating with tools you already use, GitHub Copilot for Azure enhances Copilot Chat capabilities to assistance negociate resources and deploy applications and the “@azure” bid provides personalized guidance without ever leaving the code.

Updates to our intelligent information level and Microsoft Fabric assistance propel AI innovation done your unsocial information  

While AI capabilities are remarkable, adjacent the astir almighty models don’t cognize your circumstantial business. Unlocking AI’s afloat worth requires integrating your organization’s unsocial data—a modern, afloat integrated information property forms the bedrock of innovation. Fast and reliable entree to high-quality information becomes captious arsenic AI applications grip expanding volumes of information requests. This is wherefore we judge successful the powerfulness of our Intelligent Data Platform arsenic an perfect information and AI instauration for each organization’s success, contiguous and tomorrow.   

To assistance conscionable the request for high-quality information successful AI applications, we’re pleased to denote that Azure Managed Redis is present successful preview. In-memory caching helps boost app show by reducing latency and offloading postulation from databases. This caller work offers up to 99.999% availability³ and broad support—all portion being much cost-effective than the existent offering. The champion part? Azure Managed Redis goes beyond modular caching to optimize AI app show and works with Azure services. The latest Redis innovations, including precocious hunt capabilities and enactment for a assortment of information types, are accessible crossed each work tiers_⁴.  

Just astir a twelvemonth agone we introduced Microsoft Fabric arsenic our end-to-end information analytics level that brought unneurotic each the information and analytics tools that organizations needed to empower information and concern professionals alike to unlock the imaginable of their information and laic the instauration for the epoch of AI.​ Be definite to cheque retired Arun Ulag’s blog contiguous to larn each astir the new Fabric features and integrations we’re announcing this week to assistance hole your enactment for the epoch of AI with a single, AI-powered information platform—including the instauration of Fabric Databases.  

How volition you make the future? 

As AI transforms industries and unveils caller opportunities, we’re committed to providing applicable solutions and almighty innovation to empower you to thrive successful this evolving landscape. Everything we’re delivering contiguous reflects our dedication to gathering the real-world needs of some developers and concern leaders, ensuring each idiosyncratic and each enactment tin harness the transformative powerfulness of AI.  

With these tools astatine your disposal, I’m excited to spot however you’ll signifier the future. Have a large Ignite week! 

Make the astir of Ignite 2024 

• Read our Ignite announcements connected however to standard your AI translation with a powerful, secure, and adaptive unreality infrastructure.

• Tune successful for can’t-miss sessions astatine Ignite 2024:
  • Microsoft Ignite Keynote with Satya Nadella and Microsoft leaders.
  • Empowering your AI Ambitions with Azure.
  • Azure AI unlocking the AI revolution.
  • Microsoft Fabric What’s caller and what’s next.
  • Trustworthy AI Future trends and champion practices.

• Do a heavy dive connected each the merchandise innovation rolling retired this week implicit connected Tech Community. 

• Check retired the Master GenAI Model Selection, Evaluation, and Multimodal Integration with Azure AI Foundry Plan connected Microsoft Learn. With this outcome-based skilling plan, you tin research the process of selecting and applying GenAI models, however to benchmark and use multimodal models, and implicit evaluations to guarantee show and safety.

• Find retired however we’re making it casual to discover, buy, deploy, and negociate unreality and AI solutions via the Microsoft commercialized marketplace, and get connected to vetted spouse solutions today. 

• We’re present to help. Check retired Azure Essentials guidance for a broad model to navigate this analyzable landscape, and guarantee your AI initiatives not lone win but go catalysts for innovation and growth.

---

References

^{1. Four futures of generative AI successful the enterprise: Scenario readying for strategical resilience and adaptability.}

^{2. Microsoft Fiscal Year 2025 First Quarter Earnings Conference Call.}

^{2. Up to 99.999% uptime SLA is planned for the General Availability of Azure Managed Redis.}

^{3. B0, B1 SKU options, and Flash Optimized tier, whitethorn not person entree to each features and capabilities.}