Microsoft Build is traditionally a conference centred around developers and emerging technologies. But many of the innovations showcased at this year’s event are set to reach business users across Microsoft 365, Teams, Power Platform and beyond.
From intelligent assistants that streamline daily tasks to new ways of interacting with business data, several updates introduced at Build 2025 will have a direct impact on how small and medium-sized businesses operate. Even if your business doesn’t touch development platforms, the outcomes of this year’s announcements will filter into the tools you and your team rely on every day, including productivity apps, automations and information search.
In this blog, we break down some of the most relevant updates, in plain terms, and explore how these capabilities might enhance productivity, improve decision-making and make better use of your team’s time and resources.
Microsoft introduces ‘smart agents’ to support daily work
Microsoft’s strategic investment in AI continues to gain pace, with new capabilities being developed around what it calls agentic systems, a model where AI-powered tools act more like intelligent digital colleagues. These agents are designed to understand tasks, follow structured instructions, and even collaborate with other agents to complete more complex workflows.
Although much of this development is being driven by engineers and large development teams, the benefits are now beginning to reach non-technical business users. The shift isn’t just theoretical, one of the most immediate examples is how this technology feeds directly into Microsoft 365 Copilot.
A new capability known as Copilot Tuning will allow businesses to personalise Copilot’s behaviour based on their specific workflows, business logic or industry needs. This can be achieved without any coding, through an intuitive setup in Copilot Studio. For example, a law firm might build a Copilot that applies its preferred formatting, referencing styles and tone for key legal templates, streamlining document creation across junior and senior staff. Or a consultancy could train Copilot to recognise industry-standard frameworks or client-specific terms to help surface the most relevant data or insights throughout a project lifecycle.
These agent experiences don’t operate in isolation. Microsoft is simultaneously introducing multi-agent cooperation within its AI platform. Agents can act independently or work as a team, retrieving documents, analysing inputs, and presenting summarised results for human review. This plays out as quicker responses, more context-aware suggestions, and more precise outputs across tools like Teams, Word, and Outlook.
Microsoft has also ensured that these agents follow secure practices. Tuning and interactions occur strictly within the Microsoft 365 service boundary. None of the data provided to Copilot is used to train Microsoft’s foundation models, so businesses can adopt these tools while retaining full control over sensitive content.
Copilot expands across Microsoft 365 tools
The evolution of Copilot across the Microsoft 365 suite brings a range of new, practical features that aim to eliminate manual processes and surface the right information when it’s needed most. These updates are part of the Wave 2 update, which is being gradually introduced over the coming months.
One of the highlights is enhanced support for meeting preparation. Rather than having to research, assemble documents and manually locate action items, Copilot can now prepare intelligent summaries. These include background documents, participant profiles, shared resources from previous discussions, and follow-up tasks. This not only saves time but also helps ensure that meetings are more outcome-focused.
Email productivity has also received a significant upgrade. Copilot now offers in-line summarisation, allowing users to quickly understand the content of long email threads, identify relevant attachments, and draft replies based on previous context. This is designed to support users who often spend significant portions of their time clearing inboxes or searching for specific content buried in messages.
Another important addition is Copilot Pages, an editable, shareable format that lets users turn a Copilot-generated response into a living document. These can be created directly from within a chat on mobile or desktop and later converted seamlessly into a Word document for formal presentation or distribution. Pages can contain links, summaries, interactive charts and even pre-generated content blocks, speeding up everything from meeting notes to project proposals.
Copilot’s latest updates also bring image generation capabilities through the integration of OpenAI’s GPT-4o model. This allows users to create visuals from text prompts, a useful feature for marketing, client reports, training materials and other content that benefits from visual elements.
Making data easier to use and understand
Data has never been more abundant, but making sense of business data remains a challenge, especially when reports, dashboards and analytics tools are spread across teams, files or systems. That’s why Microsoft has introduced several new capabilities that shift how users interact with data.
The headline feature is ‘Chat with your data’, which integrates directly into tools like Power BI and Microsoft 365 Copilot. This new experience allows users to ask direct, plain-language questions about their business, such as “How did our product categories perform last quarter?”, and receive real-time answers based on their existing reports and models.
This is made possible thanks to Fabric data agents, AI-driven features that understand how different data sources relate to each other and can interpret queries across organisational structures. These agents work across multiple data sources, enabling insights not just from the report you’re currently viewing but from datasets stored throughout Microsoft 365, Dynamics 365 and beyond.
Another notable improvement comes from Dataverse search, which significantly boosts the accuracy, relevance and speed of information discovery across a company’s existing documents and data. Unstructured inputs, like PDFs, scanned images or non-English sources, are converted into searchable formats, allowing users to surface information that would otherwise be missed or require specialist assistance.