Unpacking AI and Copilot Studios

Microsoft has published many Copilots within a year. We have Copilots for Azure, Security, Finance, Word, Outlook, etc. Almost all of these different Copilots are integrated into the corresponding tool. You can find Azure Copilot in the Azure Portal, Copilot for Finance in Excel and so on. The number of Copilots won’t decrease in the near future; rather, the opposite. I predict that the number of Copilots will keep growing. We still don’t have Copilot in SQL Management Studio or Copilot in Notepad. Do we need one in Notepad? Maybe not, but that is not for us to decide.

Studios

Another thing that is growing is the number of Studios. We currently have three different AI Studios: Copilot Studio, Azure AI Studio, and Azure OpenAI Studio. These Studios are tools that you can use to build your own AI solutions. Azure AI Studio is the most comprehensive and the Copilot Studio is the easiest to approach.

Copilot Studio

Copilot Studio is suitable for building your own customized Copilot in a low-code fashion. You can create your own Copilot that connects into your existing systems (SQL, Sharepoint, etc.) to get the data it uses. Tool supports multiple different publish methods for these custom Copilots. For example you can publish it as Teams Bot, or integrated it into your companies Slack channel. However, Copilot Studio lacks the deeper way to interact with Copilots features and doesn’t support a pro code model. So you are limited what the low code offers.

Azure OpenAI Studio

Azure OpenAI Studio is meant to be a tool that leverages different OpenAI capabilities and can be used to build and publish chatbots on top of OpenAI. However, it is a bit outdated, and I wouldn’t be surprised if it is deprecated and merged with Azure AI Studio soon. Currently, you can use Azure OpenAI Studio to build solutions around the Azure OpenAI service, but you can also do that in Azure AI Studio, which has a lot more features. So my current recommendation is to skip this tool and just move into Azure AI Studio.

Azure AI Studio

Azure AI Studio is a hybrid low/pro code tool. It has a low-code approach to creating an AI solution, but on the other hand, it supports Flows, which are done by using Python code. Azure AI Studio has evolved quite a lot recently and I think this is the main tool for creating AI solutions around Microsoft offering (Azure Open AI, Azure AI Search and Azure AI Services).

The home page of the Azure AI Studio says “Develop and deploy custom copilots at scale, in a safe, secure, and responsible way“, but it can do actually a lot more.  You can explore APIs and models for different use cases, build and test solutions with collaborative and responsible AI tools, safeguards, and best practices. You can also deploy AI solutions to websites, applications, and other production environments, and manage these solutions with ongoing monitoring and governance. That’s a lot more than just another copilot customizer.

Azure AI Studio is built around Hubs and Projects. Hubs are the primary top-level Azure resource for AI studio. I like to think about them as “Azure Resource Groups” that groups different AI resources and projects together. Projects on the other hand are used to organize your work and save state while building AI apps. One project should match one AI app. Some of the Hub resources like connections and compute instances are shared inside hub and you can easily use them in different projects.

Summary

Microsoft has significantly expanded its range of Copilots in the past year, integrating them into tools like Azure, PowerPoint, Word, and Outlook. This trend is expected to continue, though some tools like SQL Management Studio and Notepad still lack their own Copilots.

Microsoft has also introduced three different AI Studios: Copilot Studio, Azure AI Studio, and Azure OpenAI Studio. Copilot Studio is user-friendly for creating customized Copilots, while Azure AI Studio offers both low-code and pro-code capabilities, making it the most comprehensive tool for developing AI solutions. Azure OpenAI Studio is becoming outdated and may soon be merged with Azure AI Studio, which supports a wide range of AI development tasks, from building and testing to deploying and managing AI solutions.

As of today, Azure AI Studio stands out as the primary tool for creating AI solutions within Microsoft’s ecosystem.