This year, there has been a significant transformation in the technology sector, and two letters, A and I, accurately symbolize this shift. However, beyond these letters lies a dynamic and thrilling way in which we engage in work, communication, and collaboration. Artificial intelligence (AI) serves as a unifying element as we enter Microsoft Build, our annual flagship event for developers.
The industry has already experienced remarkable milestones this year. In January, we made a significant announcement regarding our extended partnership with OpenAI, aimed at accelerating AI advancements and ensuring their widespread accessibility. Then, in February, Microsoft introduced an entirely new Bing search engine and Edge browser powered by AI, revolutionizing the search software category, which is the largest in the world.
Here are the key advancements that have occurred at a rapid pace since then:
- Enhanced availability of Azure OpenAI Service, now offering support for ChatGPT and OpenAI’s groundbreaking GPT-4 model.
- Introduction of copilots for various user segments, such as Dynamics 365 Copilot, Microsoft 365 Copilot, and Copilot for Power Platform.
- Expansion of the new AI-powered Bing to Windows 11 taskbar, mobile devices, and Skype. Additionally, the introduction of Bing Image Creator for chat and the availability of a full open preview of the platform without the need for a waitlist.
The article emphasizes that we are merely at the early stages of the AI era, highlighting the significance of Microsoft Build. This event serves as a platform to showcase how AI is reshaping the process of building for developers and transforming the future of work.
Before delving into the news, two key concepts discussed at length during Microsoft Build are introduced: copilots and plugins.
- Copilots are applications that utilize advanced AI and large language models like GPT-4 to assist users in complex tasks. Microsoft introduced the concept of copilots nearly two years ago with GitHub Copilot, an AI pair programmer aiding developers in code writing. The company continues to release copilots across various core business domains. Copilots represent a new paradigm in AI-powered software and signify a fundamental shift in software development. They impact everything from envisioning new product scenarios to user experience, architecture, service utilization, safety, and security considerations.
- Plugins, initially introduced for ChatGPT and more recently for Bing, enhance the capabilities of AI systems by enabling interaction with APIs from other software and services. This empowers them to retrieve real-time information, incorporate business data, perform new types of computations, and safely execute actions on behalf of the user. Plugins act as the link connecting copilots with the broader digital world.
With these concepts explained, the article now directs attention towards the news and announcements being unveiled at Microsoft Build.
Expanding the ecosystem of AI plugins
Microsoft has made an announcement regarding the adoption of an open plugin standard, initially introduced by OpenAI for ChatGPT. This standard will ensure interoperability across various Microsoft copilot offerings, including ChatGPT, Bing, Dynamics 365 Copilot, and Microsoft 365 Copilot. This means that developers can use a unified platform to create plugins that can function seamlessly across both consumer and business surfaces.
Moreover, developers using Azure OpenAI Service to build their own AI applications can leverage this plugin standard, ensuring compatibility and interoperability. This empowers developers to create user experiences that allow individuals to interact with their applications using natural language, the most intuitive interface.
As part of this shared plugin platform, Bing is expanding its support for plugins. In addition to previously announced plugins for OpenTable and Wolfram Alpha, Bing will now incorporate plugins for Expedia, Instacart, Kayak, Klarna, Redfin, Zillow, and more, enriching the Bing ecosystem.
Furthermore, Microsoft is announcing that Bing will become the default search experience in ChatGPT. This integration brings a world-class search engine directly into ChatGPT, enabling users to access up-to-date answers and information from the web. The answers provided by ChatGPT will be grounded in search and web data, accompanied by citations for users to explore further. This enhanced search experience is initially available to ChatGPT Plus subscribers and will soon be accessible to free users by enabling a plugin.
Plugins are now available for developers to expand the capabilities of Microsoft 365 Copilot
We are pleased to announce that developers now have the capability to integrate their applications and services into Microsoft 365 Copilot using plugins.
These plugins for Microsoft 365 Copilot encompass ChatGPT and Bing plugins, as well as Teams message extensions and Power Platform connectors. This empowers developers to leverage their existing investments and seamlessly incorporate them into Microsoft 365 Copilot. Furthermore, developers can effortlessly create new plugins for Microsoft 365 Copilot using the Microsoft Teams Toolkit for Visual Studio Code and Visual Studio.
To enhance the functionality of Microsoft 365 Copilot, developers can also utilize the Microsoft Graph to bring in their data and provide relevant and actionable information through the recently introduced Semantic Index for Copilot.
As part of the early access program, customers will have access to over 50 plugins from partners such as Atlassian, Adobe, ServiceNow, Thomson Reuters, Moveworks, and Mural. By the time Microsoft 365 Copilot reaches general availability, thousands of plugins will be available to users.
Azure’s latest AI tooling supports developers in constructing, operationalizing, and deploying their own cutting-edge AI applications.
We are introducing the new Azure AI Studio, which simplifies the integration of external data sources into Azure OpenAI Service. Additionally, we are excited to present Azure Machine Learning prompt flow, which facilitates the construction of prompts for developers, leveraging popular open-source prompt orchestration solutions like Semantic Kernel.
Within Azure OpenAI Service, where advanced models such as ChatGPT and GPT-4 are combined with Azure’s enterprise capabilities, we are announcing updates that allow developers to deploy state-of-the-art AI models using their own data. This includes the introduction of a Provisioned Throughput SKU, providing dedicated capacity for enhanced performance. Moreover, plugins are now available to streamline the integration of external data sources into customers’ utilization of Azure OpenAI Service. Currently, we have over 4,500 customers benefiting from Azure OpenAI Service.
Developing with a shared responsibility mindset
Microsoft is deeply committed to developing AI technology that brings positive impacts and builds trust. We strive to share our knowledge and create innovative tools that assist developers and businesses in implementing responsible AI practices. During the Build event, we are introducing several significant updates.
One of the key updates is the introduction of Azure AI Content Safety, a new service within Azure AI that helps businesses establish safer online environments and communities. As part of our commitment to responsible AI systems, Azure AI Content Safety will be integrated across Microsoft products, including Azure OpenAI Service and Azure Machine Learning.
Furthermore, we are enhancing Azure Machine Learning with new tools. This includes expanding the support of Responsible AI dashboard for text and image data, which enables users to evaluate large models built with unstructured data during the model building, training, and evaluation stages. This empowers users to identify model errors, ensure fairness, and understand model explanations before deploying the models. These capabilities contribute to creating more performant and fair computer vision and natural language processing models.
Additionally, we are introducing prompt flow, soon to be in preview, which offers a streamlined experience for prompting, evaluating, and fine-tuning large language models. Users can easily create prompt workflows that connect with different language models and data sources. They can assess the quality of their workflows using measurements like groundedness to select the most suitable prompt for their specific use case. Prompt flow also integrates Azure AI Content Safety, enabling users to identify and remove harmful content seamlessly within their workflow.
Furthermore, Microsoft has announced upcoming media provenance capabilities for Microsoft Designer and Bing Image Creator. These capabilities will enable users to verify if an image or video was generated by AI. By utilizing cryptographic methods, the technology adds metadata about the origin of AI-generated content, ensuring transparency and authenticity.
Introducing Microsoft Fabric, an innovative and integrated analytics platform
In today’s data-driven world, data plays a crucial role, particularly as we enter an era dominated by AI. To harness the power of AI and create personalized AI experiences, organizations require a reliable analytics system that ensures clean data and seamless integration. However, many existing analytics systems in organizations are fragmented and lack cohesion.
Introducing Microsoft Fabric, a unified analytics platform that encompasses various essential components such as data engineering, data integration, data warehousing, data science, real-time analytics, applied observability, and business intelligence. All these components are seamlessly connected to a central data repository known as OneLake.
Microsoft Fabric provides a unified experience for customers of all technical backgrounds, enabling them to access and leverage a wide range of capabilities within a single platform. It is deeply integrated with Azure OpenAI Service at every layer, empowering customers to unlock the full potential of their data. Developers can utilize the power of generative AI through Azure OpenAI Service to discover valuable insights within their data.
With Copilot integrated into Microsoft Fabric’s data experiences, customers can utilize natural language conversations to create dataflows, generate code and functions, build machine learning models, and visualize results. Moreover, customers have the flexibility to create their own conversational language experiences by combining Azure OpenAI Service models with their data and share them as plugins.
Advancing the AI-driven future through collaborative partnerships
Our partnerships with companies like NVIDIA bring significant advantages to our customers, allowing them to create, deploy, and manage applications with the scalability and security provided by Azure. Through our collaboration, NVIDIA AI Enterprise Integration with Azure Machine Learning will enhance the development of enterprise-ready generative AI solutions. With Omniverse Cloud, exclusively available on Azure, organizations can consolidate data into large, high-performance models, integrate domain-specific software tools, and facilitate real-time collaboration across multiple factory sites. Leveraging NVIDIA GPUs, ONNX Runtime, and the Olive toolchain, organizations can accelerate the implementation of AI models without requiring extensive hardware expertise.
Enhanced functionalities for the Microsoft Dev Box
Microsoft Dev Box, an Azure service designed to provide developers with preconfigured and centrally managed dev boxes for efficient coding, is introducing several new enhancements to elevate the developer experience and increase productivity. During the preview phase, numerous customers have already experimented with Dev Box, and we have successfully migrated over 9,000 developers internally to this service for their day-to-day software development tasks.
We are now introducing additional features and capabilities, such as customization options through configuration-as-code, and the availability of new starter developer images in the Azure Marketplace. These images offer ready-to-use configurations that can be further tailored to meet specific requirements of development teams. Moreover, developers can conveniently manage custom environments using the dedicated developer portal, Azure Deployment Environments. The general availability of Dev Box is scheduled to commence in July.
Introducing Dev Home, a fresh destination for developers on Windows 11
At Microsoft Build, we are introducing Dev Home, a new Windows experience available on the Microsoft Store. Dev Home allows developers to effortlessly connect with GitHub and set up cloud development environments such as Microsoft Dev Box and GitHub Codespaces. This open-source platform is highly customizable and extensible, empowering developers to personalize their dashboard and access the necessary tools for their success. Dev Home will be available in preview during Microsoft Build.
Introducing Windows Copilot, a new feature for Windows 11
Last autumn, during our Windows and Surface launch event, Chief Product Officer Panos Panay discussed the potential of AI in revolutionizing PC interactions through features like Windows Studio Effects and DALL-E 2 in Microsoft Designer. At CES, he further emphasized the role of AI in transforming productivity on Windows.
Introducing Windows Copilot, a groundbreaking addition to the Windows platform. Windows is leading the way by centralizing AI assistance with the introduction of Windows Copilot. By integrating with Bing Chat and utilizing a variety of first- and third-party plugins, users can now focus on bringing their ideas to life, completing complex projects, and collaborating, without the need to search for, launch, and switch between multiple applications.
This development builds upon the integration of the AI-powered Bing search engine into the Windows 11 taskbar, which was introduced in February. A preview version of Windows Copilot will be gradually rolled out for Windows 11 in June.
Microsoft Build is going to be an exciting and busy event. We anticipate around 200,000 registered attendees, offering them access to 350 sessions and over 125 hours of content across two days. During the event, we will unveil more than 50 new products and features to enhance the developer experience.