Microsoft Ignite 2023: Becoming an AI-Powered organisation
Last week’s Microsoft Ignite event marked a pivotal moment in the IT industry. As a frontrunner in enterprise technology, Microsoft emphasized the integral role of Artificial Intelligence (AI) in empowering organisations and individuals alike. This evolution in technology is not just theoretical; it’s a reality that’s rapidly unfolding.
In our view, it’s imperative for organisations to develop a strategic AI roadmap. Microsoft’s announcement of over 100 updates across the entire Copilot stack is a testament to the tools and technologies available to enhance business efficiency.
As Microsoft CEO Satya Nadella said, “Copilot will be the new UI that grants access to the world’s knowledge and your organization’s knowledge. But more importantly, it’s your agent to act on that knowledge.”
He further stated, “Our vision is straightforward – we are the co-pilot company. We believe in a future where there’s a co-pilot for everything you do.”
As a company, we are extremely happy to see this direction from Microsoft because it’s very strongly aligned with our strategy of helping customers, discover, design and build AI copilots for their business and we have a bunch of AI Consulting services for this.
As you know, the Microsoft Co-pilot was generally available in early Novembers. It seamlessly integrates across various platforms, understanding your context, whether on the web, your device, or at work, and brings the right skills when you need them. You can extend Copilot’s skills by transforming your app into a plugin that increases user productivity across daily tasks and workflows.
SharePoint, the backbone of Microsoft 365, now offers SharePoint Premium. It is an AI powered content management experience integrates AI, automation, and enhanced security into content processing and governance. Features like the AI-driven “rules deviation engine” and “clause analyzer” streamline processes such as Statement of Work (SOW) preparation.
With The Teams AI Library in preview, a fast way to build message extensions in MS Teams and Copilot plugins. you can even bring OpenAI released Assistants API into your Teams applications in minutes.
AI-Powered Cloud Development:
Microsoft has also announced the preview of Copilot for Azure, accessible across the tenant level on the Azure Portal. For example, it allows for the drafting of XML based policies in API Management to boost our productivity.
In addition, the Azure Logic App Standard now features an AI powered “workflow assist” on the Azure Web Portal, aiding in workflow analysis and connector recommendations. This is a co-pilot experience you can get right now without any application for access.
At database areas, we also have a list of AI features built in Cosmos DB, such as GA of vector search for in Azure Cosmos DB for MongoDB vCore. Also, azure_ai extension to Azure Database for PostgreSQL is in public preview, which enables access to Azure OpenAI and Azure AI Language services using SQL queries with just a simple function call. You can also store, index, and query high-dimensional vectors efficiently using PostgreSQL.
Build your Own Co-pilot
There are many tools announced with similar names, so it is best we call it out together for you to understand what they are and their differences.
Azure AI Studio is now in Public Preview, with Prompt Flow and MaaS (Model as a Services). It has models other than OpenAI, including but not limited Mistral, Phi, Jais, Code Llama, NVIDIA Nemotron and more. It is a cloud based service that designed for us application developers who want to add AI capabilities to their applications without having to build and train their own models from scratch. It provides a single place to do everything you need to build an AI application, including loading data from a variety of sources, preparing it for training a model, and deploying the model to production. Azure AI Studio is ai.azure.com, please note it is different from Azure Open AI Studio (oai.azure.com)
On the other hand, Windows AI Studio is a new platform from Microsoft that aims to make it easier to develop AI apps on Windows 11. It provides a centralized place for data scientists and developers to work with all the artifacts for building, training, and deploying machine learning models. This gives us greater choice either to run models on the cloud on Azure or on the edge locally on Windows (or across the two) to meet their needs. Again, Prompt Flow makes it easier than ever to implement this hybrid AI orchestration by switching between local SLMs and cloud LLMs.
Another cool tool is Copilot Studio, formerly known as Power Virtual Agent, is a low-code tool to customize Copilot for Microsoft 365 and build our standalone copilots.
The Azure AI Content Safety is now GA. It can detect harmful user-generated and AI-generated text and images in applications and services. Azure AI Content Safety Studio is an online tool designed to handle potentially offensive, risky, or undesirable content using cutting-edge content moderation ML models.
Continuous improvements in Azure OpenAI
The first OpenAI Dev Day has so many powerful updates. For example, GPT-4 Turbo offers lower pricing, extended prompt length, and it will be in public preview in Azure OpenAI Service at the end of November 2023. GPT-4 Turbo with Vision is coming soon in Azure OpenAI for preview. DALL·E 3 is now available in public preview in Azure OpenAI Service, helping fuel the next generation of enterprise solutions along with GPT-4.
The Assistants API, including capabilities such as Code Interpreter and Retrieval, is set to enter preview soon in Azure OpenAI Service. This powerful tool simplifies the creation of agent-like experiences within applications, offering developers a versatile and efficient solution for building intelligent interfaces.
Microsoft loves SLMs too.
One of the best SLMs is Phi, a model that was built by Microsoft Research on highly specialized datasets, which can rival models and at even 50 times bigger. Phi 1.5 has only 1.3 billion parameters with state-of-the-art performance against benchmarks testing with things like common sense, language understanding and logical reasoning. In Ignite, Microsoft announced Phi 2, a scaled-up version of Phi 1.5 that shows even better capabilities across all of these benchmarks including mathematical reasoning, while staying relatively small at 2.7 billion parameters. Phi 2 is open source and can be found at Model catalog as well as models of service. With quality training data, Phi 2 can be trained to answer questions for specific domain.
Other key updates
There are of course many other major updates in key products, just to name a few below:
Microsoft Fabric is now generally available! Fabric brings together the best of Microsoft Power BI, Azure Synapse Analytics, and Azure Data Factory into a single software as a service (SAAS) platform. Fabric provides multiple workloads purpose-built for specific personas and specific tasks.
There are many new features in Azure Integration Services, such as public preview of Integration Environment and Business Process Tracking.
This year is 30 years birthday of SQL Server. For a list of exciting updates in Ignite about Azure SQL and SQL Server, we highly recommend watch this video Microsoft Ignite and PASS announcements recap – Azure SQL and SQL Server | Data Exposed
All key announcements in Microsoft Ignite can be found at this Microsoft blog.
If you are also excited about AI Transformation journey, please reach out to discuss how we can assist your AI transformation journey at https://xam.com.au/contact/.