The AI Revolution

A reasoning model does not immediately give the first answer that comes to mind. It breaks the task into steps, tries different approaches and checks its work along the way. More computing time often produces a better answer. AI can now tackle problems it could not solve before: AI systems achieved gold-medal standard at the International Mathematical Olympiad, with fully written proofs. In an organisation, a model can assess an investment under different scenarios, find a bug in thousands of lines of code or compare a new contract with all earlier agreements. A team can investigate far more scenarios and possible explanations in a short time, before people weigh the results.

Multimodal AI processes text, images, sound and video together. An assistant can answer questions about something you show it or let it hear, without everything having to be typed out. AI moves beyond the keyboard. A technician points a phone at equipment and hears what might be wrong. A conversation with a customer across the world is translated live. During an inspection, AI compares camera footage with earlier recordings and flags changes worth investigating. Personal assistants follow you between laptop and phone, remembering what you were working on. Wherever you work, AI becomes a colleague that can see and hear alongside you.

Physical AI lets machines sense, move and act in the real world. A conventional industrial robot follows a set program and repeats the same movement thousands of times. An AI robot learns from examples: videos, human demonstrations and practice in simulated environments. It can adapt when a component lies at a different angle or a box is dented. Humanoid robots, with arms, hands and human proportions, are working on assembly lines in trials. In warehouses and hospitals, robots collect and deliver goods and materials. Speed, cost and the ability to work safely alongside people still limit their use. Those are precisely the areas where progress is now rapid.

A general language model learns from vast amounts of text and knows something about many subjects. A specialist model is also adapted to one field: legislation and judgments, medical guidelines, technical manuals or an organisation's own documents. Access to those sources lets it find relevant passages and show where an answer comes from. Specialist AI goes beyond language. Models can predict molecular structures, calculate next week's weather or spot abnormalities on a scan. In organisations, these models work alongside the expert: AI handles the searching, reading and calculations; the professional weighs the evidence and makes the decision.

Generative AI creates content from instructions in everyday language. It began with text and still images. Newer models produce video with matching sound, speech in many languages and images indistinguishable from photographs. A training department makes instructional videos with spoken explanations in several languages, without a film crew. A designer shows a hundred packaging variations before any are printed. An architect walks a client through a building that does not yet exist. The next step is world models: AI that simulates an interactive environment which responds to your actions, and where robots can also practise.

An AI assistant helps with a question or task, then waits for the user to choose the next step. An AI agent receives a goal and takes the steps to reach it: finding information, opening a system, filling in a form, sending a message and checking the result. An agent prepares an entire application from start to finish, instead of writing a single piece of text. AI agents are improving rapidly at real computer tasks and handling longer sequences of work. The Agentic AI keynote examines what this means for work, oversight and the limits of what an AI agent is allowed to do independently.

Many organisations add AI to existing work: an email assistant, a report after a meeting. An AI-first organisation starts from the other end. AI and new technology do the work people need not do, including the repetitive and predictive work, and people focus on the work where they make the difference: judging, choosing and working with others. At a housing association, a repair request moves from the initial report to a booked technician without intervention, leaving a housing officer more time for a resident who needs help. McKinsey research finds a strong link between redesigning workflows in this way and the financial results organisations attribute to AI.

An AI tool can be available quickly; the work around it changes more slowly. Behaviour often explains the gap. Familiar habits, fear of becoming redundant, a sense that expertise is losing value, a lack of time to practise and uncertainty about what is allowed are common across many organisations. KPMG reports that many employees already use AI, though often without saying so. Once that use becomes visible, adoption often accelerates. A colleague showing how AI saved an afternoon makes the possibilities concrete for the whole team. Early adopters help AI use grow from the ground up, until an occasional experiment becomes a habit.

When AI analyses scenarios, predicts deviations in processes and shows decision-makers new connections, people's work shifts. People carry out fewer tasks themselves and spend more time directing AI: setting tasks, assessing results and intervening where the system lacks context. New roles emerge, such as checking AI's work or designing processes around it. In a human-AI team, AI prepares the work; people bring judgement, human contact and accountability. Entry-level work changes too. Finding sources and organising case files once helped newcomers build expertise. AI increasingly takes on those tasks, so new ways of learning the profession are emerging for people at the start of their careers.

Employees often judge where an organisation is heading by how its leaders behave, rather than by its written strategy. Leaders who set the example take the initiative: they use AI themselves, show what it delivers and make room to learn and try things out. The future then becomes something people build together. That room has limits: where mistakes have serious consequences, practice begins in a safe environment. The leader's own work changes too. AI brings more scenarios and signals into view. Decisions about direction, people and values remain the responsibility of those leading the organisation.

AI literacy means understanding AI's capabilities and limits, how to work with it, where answers can go wrong and which data may be entered. Recognising a convincing answer is different from being able to judge what it is based on. For an employee, this might mean understanding a prediction; for a leader, examining the assumptions behind a scenario. The EU AI Act requires organisations using AI to take measures that support the AI literacy of people working with it. A future academy, where tomorrow's employees and leaders develop their skills, allows that understanding to grow alongside their professional knowledge.

AI is developed by major model builders, specialist companies, start-ups and research groups worldwide. The best ideas often come from peers who have already taken the next step. An AI ecosystem is the network through which an organisation shares knowledge, data and solutions. Hospitals jointly train a model while keeping patient data within their own institutions, allowing it to learn from many institutions at once. Transport operators share planning data to avoid empty runs. As development accelerates, that network matters more: work already tested elsewhere does not have to be reinvented.