Designs for the frontier future
Source: Jon Friedman; Microsoft Design · 2025-05-30
At its core, this reimagination sets the stage for a new era of intent-driven computing where systems begin mapping to how people think, feel, and act, rather than the other way around. It also both reflects and propels the dawn of human-AI collaboration at scale, where individuals become “agent bosses,” empowered to direct intelligent systems toward their goals with clarity and purpose.
Beyond empowering customers in new ways, it also marks a meaningful shift in how Design helps build AI-first experiences. As UX professionals move from creating interfaces to orchestrating coherence across complex and nondeterministic systems, we play a heightened role in turning clutter into clarity, fragmentation into flow, and intention into impact.
Before diving into the M365 Copilot experience itself, let’s start from where our design process always begins: a 30,000-foot view that considers the world we’re designing for. What do people need today? What will they need tomorrow? How can we create not just features, but futures?
Apparent in our new M365 Copilot app is the next iteration of Fluent as a design system to support this professional shift. The future of work is less about static roles and more about dynamic goals and how you get things done, fueling a demand for unified, fluid experiences that foster collaboration between people and AI agents. As I wrote earlier in Design systems for the AI era, our systems must begin focusing more on verbs and action above static icons and move from passive UIs to adaptive systems. As controls become more intelligent, inputs more flexible, and outputs more personalized, we’re not just designing products but shaping a digital workforce. One where you are in the driver’s seat, conducting an orchestra of Copilot agents to make your ideas a reality.
For decades, design helped people grasp technology by translating computing complexity into something approachable and human. Metaphors like desktops, folders, and the cloud helped bridge the gap between abstract systems and everyday understanding. Today, technology is beginning to adapt to us, learning our behaviors, understanding our goals, and responding to our intent. The burden of translation is shifting, and systems will begin meeting us on our own terms—human terms.
The new Microsoft 365 Copilot app is built on a simple, powerful truth: conversation is core to how we understand and relate to the world. Long before we speak, we begin to absorb the cadence and emotion of language, even in the womb. That early exposure shapes how we connect, learn, and collaborate. Copilot Chat builds on this deep human foundation, anchoring the experience in natural, intuitive dialogue between people and AI. With Chat as the primary interface for work, people can initiate, navigate, and complete work within a single, language-driven environment, instead of jumping between siloed tools. It’s a paradigm shift from app-centric workflows to intention-centric ones and a turning point in how we interact with productivity tools.
Copilot’s dynamic windowing model further enhances the experience, paving the way for more adaptive interfaces. Users can move fluidly between their task and AI collaboration, whether it’s a side-by-side chat beside your Pages document or editing tools served up in the moment to refine an AI-generated image.
This isn’t about baking AI into interactions for the sake of it but considering where intelligence belongs and why. It’s about choosing the right agents for the right jobs and empowering people to drive the outcomes that matter to them, pushing productivity to new heights. By designing a UI that flexes with user intent, we’re setting the stage for a future where AI doesn’t just assist but collaborates, anticipates, and adapts in real time. With added memory and a growing agent platform supporting first- and third-party integrations, Microsoft 365 Copilot is evolving into a system that brings the right tools, context, and insights when and where they’re needed.
Visually and structurally, we also made deliberate decisions to support clarity and reduce cognitive load, something WTI showed to be a top concern. Microsoft 365 Copilot’s interface is built on an accordion-style architecture, with a collapsible navigation pane and a clean, minimal palette. In a world where constant interruptions define the modern workday, it brings calm and restores focus, creating space for what matters most.
As generative AI reduces the cost of creation, the role of Design is shifting from making things to making sense of things. We’re no longer just builders. We are guides, bringing editorial clarity to increasingly complex workflows, especially in an AI-driven world where intelligence is no longer a limited asset but an on-demand resource.
Design has become a connective tissue across the Microsoft ecosystem, coordinating how AI agents operate across different contexts to deliver insight, continuity, and coherence. At the same time, because we are rooted deeply in lived experiences and user research, we bring an empathetic lens to our processes and decisions, championing important principles around accessibility, security and responsible AI.
These aren’t abstract ideals. They’re reflected in how Copilot is designed and shows up in the world. Through regular security audits, UX guidelines, documentation and resources, we ensured that all new libraries, prototypes and experiments were secure by default while we built the new app experience. We also tightly integrated our Responsible AI guidance across our UX themes for all platforms. Copilot also builds on Microsoft’s longstanding investments in inclusive design; features like speech-to-text, image captioning, and natural language input make it easier for people of all abilities to engage.
These choices aren’t just about functionality—they’re about trust, inclusion, and doing right by the people we design for.
Lines of inquiry this paper opens 24
Research framings built by reading the notes related to this paper — the questions it feeds into.
When do generative or structured interfaces outperform conversational chat? How do users build accurate mental models when agent reasoning is partially hidden? How should control and autonomy be allocated in human-AI workplace collaboration?- How do designers decide when AI agents should escalate to humans?
- Can AI systems make their assigned co-pilot properties transparent to users?
- How does the agent boss model differ from traditional AI copilot metaphors?
- Do self-selected heavy Copilot users show different task shifts than randomized users?
- Who decides what work a coworker agent should do daily?
- What capabilities must agents have to support human-centered service delivery?
- How do autonomy levels translate the co-pilot metaphor into practice?
- How does agent design differ fundamentally from interaction design practice?
- Where should AI integrate to avoid creating isolated product experiences?
- Does agent experience design produce better outcomes than screen-first design?
- How do role dynamics change when UX designers, engineers, and product managers prototype together?
- Do just-in-time objectives require active designer judgment at every interaction?
- Why should service design absorb agent design rather than keep them separate?
- What evidence would prove agent design has achieved paradigm status?