Tables and pages have been cornerstones of human knowledge organisation for millennia; the first datable mathematical table comes from the Sumerian city of Shuruppag circa 2600 BCE. This week, Microsoft unveiled a wave (their so-called “wave 2”) of AI updates to its Copilot suite, including a new collaborative canvas called Copilot Pages, and a big AI upgrade to the venerable Excel.
The Pages tool aims to bridge the gap between AI-powered chat interfaces and traditional office docs, although the danger of more screens to deal with in Office 365 is a real one. At the same time, Microsoft announced the general availability of Python in Excel, leveraging AI chat, thus democratising real data science for the masses.
Pages represents a new idea for how users might interact with AI assistants. Rather than confining interactions to chat windows or embedding them within existing document types, Pages creates a new, flexible space where teams can collaboratively work with AI. This approach could address some of the limitations users have experienced with Copilot (despite much fanfare it has been a big disappointment in most of its guises), allowing for ‘richer’ interactions.
“This is an entirely new work pattern — multiplayer, human to AI to human collaboration,” says Jared Spataro, corporate vice president of AI at work at Microsoft. The ability to share Copilot Pages with a simple link and embed them in other documents shows Microsoft is thinking about how to integrate AI more seamlessly into existing workflows. It needs to, as this is perhaps one of the biggest barriers to mainstream business adoption today. We still mostly work in the old documents, emails and the team-based chat windows of the past.
Alongside Pages, Microsoft is pushing forward with AI agents, customised bots with specific skills that can handle some tasks without direct user input. This move aligns with a broader industry trend towards more proactive AI systems, as seen in recent announcements from Salesforce. However, Microsoft’s approach of positioning agents as an extension of Copilot, rather than a replacement, raises questions about how businesses will manage an increasingly complex ecosystem. Microsoft’s simple Copilot agents can be created to have access to particular knowledge bases, documents, or data sources, and they can be given specific instructions or prompts to guide their behaviour, much like OpenAI’s custom GPTs. This allows them to provide more focused and relevant assistance in particular contexts. For example, an HR agent might be created with access to company policies and procedures, while a sales agent might be loaded with product information and customer data.
The integration of Python into Excel could have just as many implications for businesses. By bringing Python’s powerful libraries directly into the familiar Excel environment, with the help of easy-to-use chat (now leveraging GPT-4o which is a much-needed upgrade), Microsoft is dramatically lowering the barrier to entry for advanced analytics. Peter Wang, Co-Founder and Chief AI & Innovation Officer at Anaconda, describes it as “a major breakthrough that will transform the workflow of millions of Excel users around the world.” And that could mean millions of businesspeople given the continued prevalence of the Excel table as the central construct for business.
Takeaways: Pages and tables have stood the test of time for a reason, and if enhanced correctly, could be the next paradigm for AI, integrating the segregated islands of chat, email, documents, and files. After some early experiments and lacklustre tools, the new Copilot, plus Pages, Agents, and Python in Excel form the real shape of the Microsoft’s future end-user experience. Integration and access to data will likely be underwhelming at first, and one-day agents will do most of the work, but meanwhile unlocking adoption will be key. Every Microsoft using business should be looking to form one or more superuser squads to really test and explore these new features in context. This is not just the roll-out of a new version upgrade, or a background move to the cloud, this is the future of productivity.