Over the past few months Microsoft has announced a range of first-party Copilot applications for Microsoft 365, Dynamics 365, Power Platform, and more. Today, we’re sharing the app development pattern we used to create these solutions—and showing how any developer can use this Copilot stack to develop their own copilots. What’s more, we’re excited to announce a joint commitment with OpenAI to support and grow the AI plugins ecosystem, embracing an open standard that will enable plugins interoperability across ChatGPT and the breadth of Microsoft’s Copilot offerings.
Developers can build plugins that work across both consumer and business surfaces, including ChatGPT, Bing Chat, Dynamics 365 Copilot, Microsoft 365 Copilot, and other Microsoft first-party Copilot apps. This means developers can build experiences that enable people to interact with their apps using the most natural user interface: human language..
Build next-gen, AI powered applications
We believe AI is the next massive shift in computing. That’s why at Microsoft we’re investing heavily in AI and the tools to deploy AI, so developers and organizations can do more. With the latest innovation announced at Build, developers can create unique experiences using comprehensive developer tools and new AI and machine learning capabilities. To learn more, read the blog post by Jessica Hawk, CVP of Data, AI, and Digital Applications Marketing.
Our latest Azure AI and Machine Learning innovations include the following:
- We’re proud to announce Azure AI Studio and new capabilities to harness generative AI, including the capability to easily ground OpenAI models on your data, coming to preview. With just a few clicks, developers can now ground OpenAI models, such as ChatGPT and GPT-4, on their data to quickly and easily to build organization specific conversational AI experiences.
- Azure AI prompt flow coming to preview, will provide a streamlined experience for prompting, evaluating, tuning, and operationalizing large language models. Developers and data scientists can quickly create prompt workflows that connect to hundreds of popular open source and proprietary models and data sources for building intelligent applications and assessing the quality of their workflows to choose the best prompt for their use case.
- Azure OpenAI Plugins, coming to private preview, will streamline the process of building and consuming APIs that extend the capabilities of GPT-4. The following plugins will be available during private preview: Azure Cognitive Search, Azure SQL, Azure Cosmos DB, Microsoft Translator, and Bing Search.
- To make working with open-source models easier, we’re also introducing foundation models in Azure Machine Learning, which starts with the model catalog to select from collections of foundation models including both Azure OpenAI Service models, and open-source models curated by Azure Machine Learning and Hugging Face, and provides the ability to fine-tune and deploy those foundation models using Azure Machine Learning components and pipelines.
- New Vector search capability in Azure Cognitive Search, in private preview, allows users to now store, index, and search within their datasets based on vector representations of their data, also known as embeddings, to find information that is semantically similar to their search query. Vector search can be used in combination with retrieval plugins for ChatGPT through Azure OpenAI service.
- We are now enabling a Provisioned Throughput Model for Azure OpenAI Service, to offer dedicated capacity.
- New Azure AI Content Safety service will make it easier for developers to test and evaluate AI deployments for safety by detecting and assigning severity scores to unsafe content across languages in both images and text. We’re integrating Azure AI Content Safety across products, including Azure OpenAI Service and Azure AI Studio, and Azure Machine Learning, to help practitioners assess models prior to deployment and as a content moderation tool.
We also announced exciting updates to our comprehensive developer tools and app platform portfolio, including:
- Microsoft Dev Box, generally available in July, is an Azure service that gives developers access to ready-to-code, project-specific dev boxes that are secure and centrally managed. Microsoft Dev Box helps support hybrid dev teams of any size, helping developers focus on writing code by streamlining access to all the resources and tools they need for the project at hand.
- GitHub Advanced Security for Azure DevOps in preview soon, is a solution that provides the three core features of GitHub Advanced Security into the Azure DevOps platform, so customers can integrate automated security checks into their workflow. It includes code scanning powered by CodeQL to detect vulnerabilities, secret scanning to prevent the inclusion of sensitive information in code repositories, and dependency scanning to identify vulnerabilities in open-source dependencies and provide update alerts.
- Azure Deployment Environments (ADE), now generally available, enables developer teams to quickly spin up app infrastructure with project-based templates, minimizing setup time while maximizing security, compliance, and cost efficiency. ADE provides self-service templates that deploy directly from dev tools, code repos, or custom developer portals, and maximize security with centralized permissions and policy governance, access controls, and full management of cloud resource configurations.
- Azure Kubernetes Service (AKS): To give enterprises more control over their environment, we are announcing long-term Support for Kubernetes that will enable customers to stay on the same release for two years—twice as long as what’s possible today. We are also announcing confidential containers in AKS, coming soon in preview, as a first party offer that allows teams to run standard unmodified containers, aligned with Kata Confidential Containers opensource project.
