For the past forty years, the application has been one of the fundamental units of work. We would log onto a desktop computer, laptop or mobile computer, open apps, perform tasks, and close it. Our relationship with computing has been defined by the ‘canvas’ of the graphical user interface (GUI), and it has served us well. However, with the advent of agentic AI, we will witness a new and powerful and fluid paradigm: the composable user interface.
This isn’t a simple evolution. It’s a fundamental restructuring of human-computer interaction, driven by the convergence of four powerful forces:
· Awareness of the user’s skills and knowledge
· contextual awareness of ambient surroundings
· reasoning power of multimodal AI
· connective tissue of the API economy.
For the enterprise, and particularly for the frontline worker, this shift means technology will no longer be a destination, but a pervasive, intelligent layer that configures itself around the frontline worker and the task at hand. Our daily work is increasingly shaped by conversational, agentic AI and its fundamentally rewriting how we get things done – creating new ways of working.
Apps are not going away, but in the longer-term future, we will engage an AI-orchestrated collection of capabilities that manifest as a dynamic interface, assembled in real time uniquely for the connected frontline and environment.
Ambient Surroundings: The Fabric of Context
To deliver this personalised and connected frontline future, agentic AI requires a rich, continuous stream of data. This is where the concept of a “fabric of sensing” (the foundation of true ambient computing) comes into play. Imagine a world where a multitude of sensors, cameras, RFID tags, GPS, temperature, motion, barcode and other data capture points on the frontline work in concert to create a real-time, digital twin of an environment.
This “always-on” sensing provides deep contextual data that allows AI agents to understand the world with unprecedented granularity. It provides intricate details like motion, location, device health, temperature, task list, location of colleagues, asset and inventory. It even delivers the user’s current state to enable these agents to move beyond simple commands and begin to anticipate needs, proactively offering assistance and insights.
It’s this fabric of data that transforms the AI from a passive tool into an active collaborator, moving from “what is happening” toward where, when, and why things are happening and importantly what comes next.
The Composable UI: An Interface Built for the Moment
A striking outcome of this shift is that the user interface itself will become spontaneous. Instead of a one-size-fits-all design, the AI will generate a composable UI on the fly, specifically tailored to the immediate task.
Consider the implications. A retail associate’s interface might dynamically reconfigure to show customer loyalty information when they are on the shop floor talking to a customer but then morph to display stock-check and location functions when they enter the stockroom.
The composable UI of the future won’t have a single, fixed interface. Rather, it will transform based on worker activity, like a retail assistant moving between handling tasks in a back of store stockroom to shelf replenishment to talking with a customer about an order.
This level of personalisation will be driven by learnings from thousands of users in similar roles, creating a constantly evolving and hyper-efficient user experience. The interface is no longer a static canvas but a dynamic partner, adapting its form to best suit the function of the moment.
A New Paradigm for Work: Orchestration and Augmentation
The future of work will be collaborative in new ways between humans and their digital assistants. We will see the emergence of both personal and enterprise agents working in harmony to augment the capabilities of the frontline worker.
A task that begins on a handheld mobile computer with AI agents on-device could automatically transfer to an in-truck display when a worker steps onto a forklift, with the AI orchestrating the experience across devices to ensure both safety and efficiency.
In this model, the user carries their “digital brain” with them, and the environment provides the screens. The AI projects or switches to the most appropriate interface onto the most convenient surface, whether it’s a wearable, a desktop computer, a fixed terminal kiosk, or even an audio cue in an earpiece. This frees the worker from being tied to screens, allowing them to focus on the physical world while being guided by a hands-free, intelligent assistant.
This is about efficiency and augmenting human capability. In healthcare, nurses spend a phenomenal amount of their time on documentation. In the future, a nurse equipped with a healthcare mobile computer to assist with bedside tasks is may be complemented by an AI-powered environment, and could simply state, “The patient in this room seems comfortable.”
The AI, fed by the fabric of sensing, would automatically layer in the vital signs from monitors, confirm medication schedules, and log the entire interaction for the nurse’s final review and approval, and remains accessible on the mobile computer in the pocket. The clinician becomes a “human-on-the-loop,” freed from manual data entry to focus on what truly matters: patient care.
Of course, this vision is not without its hurdles. An “always-on” fabric of sensing requires potential data security and privacy issues to be addressed with robust governance, cybersecurity, and transparent policies. Acknowledging these challenges is not a barrier to adoption but a prerequisite for it. They can be solved with the same ingenuity we apply to technology itself.
The Disruption of Capabilities over Applications
This profound shift demands a new mindset, and nowhere is the disruption more acute than in software development. For decades, Independent Software Vendors (ISVs) and internal IT departments have built and sold applications. In the new world, their value will come from creating discrete “capabilities” that can be plugged into an AI agent.
This could fundamentally change the traditional software development lifecycle. The focus moves from building applications to engineering modular, API-first capabilities – a specific inventory check, a patient data query, a machine diagnostic that an overarching AI can then weave together. A developer’s customer is no longer only a human user, but the agentic AI that serves that user.
This transition is inevitable. The question for today’s tech leaders is no longer “Which apps should we build?” but rather, “What unique capabilities must we develop for customers to remain relevant in an AI-first, composable world?” The answer will determine who thrives in this new era and who becomes a relic of the one that is rapidly disappearing behind us.
