Microsoft has officially moved on from its chatbot experiment, instead focusing on the development of agentic AI for business process automation. At Microsoft Ignite 2024, the company showcased its vision for using large language models (LLMs) to deliver self-assembling process automations that can convert natural language queries into long transactions across well-defined service endpoints.

This shift in focus is not surprising, given the limited productivity gains from LLM-powered chatbots. Microsoft is now leveraging its experience with Copilots to develop new tools that can be used to drive tangible productivity gains. The company's Microsoft 365 Copilot and Semantic Kernel are key components of this strategy, enabling users to work with OpenAPI-powered services and build low-code agents.

The concept of agentic AI builds upon the 30-year-old idea of autonomous agents, combining various LLM-based development techniques to deliver self-assembling process automations. Microsoft's Azure AI Foundry is driving this shift, and the company is now putting these tools in the hands of low- and no-code developers through Copilot Studio. The goal is to kickstart agent development, not from traditional developers, but from business users who are at the heart of complex business processes and are looking for tools to simplify things.

Copilot Studio features, such as the ability to build low-code agents, were detailed at Ignite 2024, providing context and important details on how low-code development will integrate with more complex AI applications built with Azure AI Foundry. This fusion of custom and prepackaged AI fits with Microsoft's fusion teams model, allowing users to build and tune custom multimodal AI models that can be quickly included in their own applications.

Power Platform, which has historically stood apart from Microsoft's stable of developer tools, is now being integrated into the full developer ecosystem. This change is driven by Microsoft's understanding that process automation requires significant input from both business and technology experts. Agentic AI will require both programming and process-modeling skills to build out a pipeline of AI assistants supporting a long workflow.

The Microsoft 365 Agent SDK, announced at Ignite, provides a set of tools for building Microsoft 365-based agents, allowing access to Copilot Studio applications from C# code, with Python and Node.js support on the roadmap. This SDK bridges not only Azure AI Foundry and Copilot Studio but also enables work with third-party AI platforms.

The Microsoft 365 Agent SDK has four key components, offering multiple user interfaces, grounding data from services like the Microsoft Graph, Azure Fabric, and Azure AI Search, and orchestration from Semantic Kernel. Integration between Copilot Studio and the Microsoft 365 Agent SDK is two-way, allowing users to add memory and skills to a Copilot Agent or call an existing Copilot Agent without needing to build API integration.

This mix of low code and pro code should give both platforms flexibility, enabling users to mix and match features to build and deliver the appropriate agent-powered workflow for their specific business process needs. As the technology continues to evolve, it will be interesting to see how tools like these step outside traditional workflows, potentially controlling user PCs and automating workflows that extend from the data center to desktops.

Microsoft's focus on agentic AI for business process automation marks a pragmatic approach to delivering new technology. As new tools and models arrive, the company's vision is likely to adapt and change quickly. With its focus on collaboration between software developers and users, Microsoft is providing different ways to mix low-code tools with traditional development platforms, giving users the ability to choose the approach that's right for them, their experience with AI, and their business needs.