AI + Machine Learning, Azure Cognitive Search, Azure Cosmos DB, Azure Machine Learning, Azure OpenAI Service, Events, Microsoft Purview, Text Analytics
What’s new in Azure Data & AI: Helping organizations manage the data deluge
By Jessica Hawk Corporate Vice President Data, AI, and Digital Application Product Marketing
7 min read
When I meet with customers and partners, one thing is clear: nearly every organization is looking to accelerate its use of data to drive strategic outcomes. Generating data isn’t the issue. By 2025, IDC estimates the world will surpass 181 zettabytes globally and a staggering 80 percent of that data will be unstructured (think streaming audio and video, social media, and the Internet of Things (IoT) sensor data).1 This explosion of data presents both an opportunity and a challenge. The opportunity is that we can use this data to uncover as-yet unknown insights, predictions, and automation opportunities to drive business outcomes in previously unthinkable ways. Data fuels AI innovation, which is something many of us are being asked to do given the recent advances in generative AI. However, we as practitioners understand that a key challenge lies in securely ingesting, storing, managing, analyzing, and sharing all that data effectively.
These challenges and opportunities that come as a result of today’s exponential data proliferation exist across industries. For example, the National Basketball Association (NBA) now ingests and analyzes around 10 million data points per game to influence strategies on the court. SEGES, a non-profit that advises farmers on running their businesses more efficiently and sustainably, uses big data to predict crop yields and spot early signs of livestock disease or injury with higher accuracy. AG Insurance relies on vast real-time data to build more personalized insurance solutions for customers. While each organization has distinct use cases, they all use the Microsoft Intelligent Data Platform to modernize their data estate and enable advanced analytics and AI. We see this over and over with our customers—getting the data house in order is truly foundational to taking advantage of what generative AI can do for a business. To get expert help with designing and building a modern data foundation for AI, check out the Azure Migration and Modernization Program.
Yet we know getting to a modern, cloud-based data foundation doesn’t end with a shift to the cloud. Data governance, security, and privacy are fundamental to a modern data estate. In March 2023, Microsoft became the first cloud provider to successfully complete the Cloud Data Management Capabilities (CDMC) 14 Key Controls and Automations certification. This certification demonstrates Microsoft’s commitment to providing comprehensive cloud data management automation and controls for protecting sensitive data to accelerate cloud adoption. We are excited to continue to partner with leaders around the globe to make data governance in the cloud easier to implement and more approachable which will allow everyone to derive more value from data responsibly.
Hungry to hear from industry peers as they discuss their approach to data strategy? Check out Insights Tomorrow, a new podcast from Microsoft hosted by Patrick LeBlanc featuring in-depth conversations with data leaders and experts about the revolutionary journeys they are taking in the world of data, analytics, and governance.
New and upcoming technical resources:
- Microsoft Build: Join the next wave of developer innovation. Registration is now open for in-person and virtual experiences May 22 to 25, 2023. Get the latest product news from Microsoft, advance your knowledge and skills with deep interactive sessions, and network with peers.
- Can ChatGPT work with your enterprise data? In 15 minutes, learn how to integrate ChatGPT into your own enterprise-grade app experiences using Azure OpenAI Service with precise control over the knowledge base, for in-context and relevant responses. Interact with your organization’s private internal data, while respecting the information protection controls put in place. Watch “Can ChatGPT work with your enterprise data?” to learn more.
- Learn Azure data fundamentals: Kickstart your learning journey with our new Cloud Skills Challenge. In 30 days, you’ll gain a foundation in core database concepts in cloud environments and data services including relational data, nonrelational data, big data, and analytics.
We’re innovating across the Microsoft Cloud to make data more accessible and actionable for every line of business. This month, we are introducing new capabilities to help you move beyond the cost and complexity of fragmented data to realize business value with AI. Let’s dive in.
Start small and scale relational apps globally with Azure Cosmos DB for PostgreSQL
Azure is the first cloud provider to offer a single database service that supports both relational and non-relational workloads, helping you build cloud-native apps with all your data. Azure Cosmos DB for PostgreSQL is a fully managed service, enabling highly scalable relational apps, such as multi-tenant software-as-a-service (SaaS) apps, real-time operational analytics apps, and high-throughput transactional apps. This month, we announced the general availability of cluster compute start and stop to easily stop and start clusters as needed to control compute costs and REST APIs to help customers manage cluster and networking settings programmatically. We also announced the preview of Customer Managed Keys, enabling you to bring your own key for data protection at rest in Azure Cosmos DB for PostgreSQL. Learn more about Customer Managed keys on the Azure Cosmos DB blog.
Do more with less using Centrally Managed Azure Hybrid Benefit for SQL Server
Azure Hybrid Benefit is a cost-saving benefit that allows customers to apply their existing SQL Server licenses with Software Assurance to Azure virtual machines, saving up to 30 percent or more on Azure SQL Database and Azure SQL Managed Instance. To help alleviate the financial pressures organizations are facing amidst economic uncertainty, we are pleased to announce the general availability of Centrally Managed Azure Hybrid Benefit for SQL Server, which helps customers optimize costs by applying blocks of eligible licenses across all Azure SQL virtual machines and platform-as-a-service (PaaS) resources within their overall account or subscriptions. With full visibility into a SQL Server estate, centralized management aims to make it easier for customers to monitor and sustain compliance and get the most out of their SQL investments.
Operationalize responsible AI for healthcare applications
When making real-world decisions that impact patients, practitioners, and facilities, it is critical that healthcare institutions leverage the power of AI and machine learning models responsibly. At the HIMMSS23 conference earlier this month, we announced the general availability of the Azure Machine Learning responsible AI dashboard accelerator kit for healthcare. The responsible AI dashboard is helpful across industries, but this accelerator kit aims to help machine learning practitioners train and debug models specifically for healthcare settings. For example, this accelerator can equip healthcare providers with directional and causal relationship analyses between historical patient data and health scores to help inform lifestyle recommendations and modifications for patients. By prioritizing model fairness and explainability, healthcare organizations can enable higher confidence in and adoption of AI outputs to support patient care, space planning, or dispatching and scheduling staff.
To tie it all together, I’ll also share an abbreviated demo from the showcase floor at HIMMS23. Below, we see how Contoso Healthcare, a nationwide healthcare provider, leverages the Microsoft Intelligent Data Platform to maximize the value of their data estate to help them reduce costs, improve clinical operations, and manage their organization’s compliance requirements. We get a look at how vertical AI solutions, such as Custom Text Analytics for healthcare, can help physicians save time when reviewing their patient’s medical history, and we see how Azure Cognitive Search and Azure OpenAI Service can help physicians find the answers they need across vast quantities of medical research documents more quickly. Next, Contoso combines the Azure Machine Learning responsible AI dashboard for healthcare and Power BI to help their physicians understand patterns in their data to improve clinical and operational insights. Lastly, Contoso uses Compliance Manager, part of Microsoft Purview, to help them manage their organization’s multi cloud compliance requirements.
Identify up to 10 spoken languages in real-time with Azure Speech
We are proud to introduce Continuous Language ID, our newest feature available with Azure Speech services. With this feature, speech recognition and speech translation customers can now seamlessly detect multiple languages in real-time scenarios—such as live presentations or broadcasts—as well as analyze existing audio content.
Continuous Language ID builds upon existing speech identification capabilities available in Azure Speech services that can detect a single language within the first five seconds of a piece of audio. Now, with Continuous Language identification, users can accurately detect up to 10 languages throughout an entire piece of audio content enabling them to transcribe or translate multi-lingual audio content in one go. If you anticipate multiple languages will be spoken in real-time scenarios or want to analyze existing audio where speakers use different languages, Continuous Language ID is the feature you need. You can try Azure Speech services for free today.
Train state of the art models with Azure AI Infrastructure
To close, I want to share new resources to learn about how Azure’s leading cloud AI supercomputing infrastructure and end-to-end machine learning capabilities give you the performance, scalability, and built-in security needed to build, train, and deploy the most demanding AI workloads with confidence. From the training of state-of-the-art complex AI models to deep learning, and inference, Azure delivers solutions backed by Microsoft’s responsible AI principles, and consistency for data scientists with a familiar toolchain. Learn about what makes our infrastructure differentiated and our proven track record of delivering the greatest supercomputing to partners like OpenAI and Hugging Face to build and host their own large AI solutions with this infographic and the new whitepaper from Harvard Business Review, “Rethinking Cloud Strategies for Advanced AI.”
Start your data modernization journey with Microsoft
Whether you have large, complex AI workloads or you simply want to make the most of intelligent capabilities in your existing applications, Microsoft and our partners are ready to help you accelerate your productivity with data and AI. If you missed it last month, I recommend checking out our webinar with Forrester, “Accelerate AI In Your Org—It’s Easier Than You Think” to learn more about the impact of AI on the workforce, the challenges of AI adoption, and the potential of AI to drive innovation and growth. And of course, I truly hope you will join us at Microsoft Build on May 23 to 25, 2023 to learn about exciting innovations, get hands-on experience, and network with peers.
1 IDC Corporate, “Worldwide IDC Global DataSphere Forecast, 2021-2025”, March 2021.