Surprise insight: Mid-career rethink BOX 6 Most public debates assume that AI’s biggest impact will fall on entry-level workers, as routine starter tasks are automated. Yet survey responses from C&T firms suggest a different possibility: the real disruption could emerge in the middle ranks. –If junior staff can ramp up faster with AI co-pilots and academies, the need for long apprenticeships diminishes. –If specialists can concentrate directly on higher-order work, coordination layers become less critical. –That leaves mid-career professionals – whose value often lies in supervising, coordinating or gatekeeping information – potentially more vulnerable than either juniors or senior experts. This theme was voiced cautiously and without hard numbers, but it emerged often enough to warrant attention. AI may not just reshape how jobs start or finish – it could quietly hollow out the middle of organizational hierarchies. At the specialist level, AI is already shouldering pieces of judgement-heavy work. Firms described its use in contract review, compliance reporting and even design iteration (see Box 1). The mood here was not defensive but pragmatic: professionals welcomed AI’s ability to strip out time-consuming, detail-intensive work, freeing them for negotiation, oversight and client engagement. Still, the fact that these tasks are automatable challenges assumptions about what kinds of expertise are uniquely human. Taken together, these dynamics suggest that AI is reshaping job hierarchies less through wholesale elimination than by rearranging the balance of roles across tiers: reimagining the entry stage, pressuring the middle and refocusing specialists on the highest-value dimensions of their work. Today’s AI adoption is uneven. Large enterprises in developed markets dominate the early use cases, while adoption in smaller firms, governments and organizations in emerging markets has been non-linear, with both major breakthroughs and considerable caution. Executives saw this not as a story of lag, but of potential. For smaller firms and non-profits, out-of-the-box tools and “AI-as-a-service” offerings allow lean teams to automate routine work, stretch tight budgets and deliver more personalized services. Several leaders described this as a way for small organizations to “punch above their weight”. They also provide opportunities for so-called AI-First Enterprises: AI-native start-ups that are pushing the boundaries of innovation with tiny teams by fully embracing AI tools from day one. In emerging markets, executives pointed to leapfrog opportunities: AI-native solutions accessed principally via smartphones that bypass legacy infrastructure in fields such as healthcare, education and citizen services. At the same time, technology leaders flagged sovereignty concerns – reliance on a few global providers could leave countries or smaller players vulnerable. The adoption and promotion of interoperability will have a significant impact on facilitating compatibility and integrating AI models 2.4 Uneven adoptionIn practice: The Cisco Networking Academy’s Data Science Essentials with Python course delivers project- and problem-based learning at scale. Rather than passively consuming static content and then completing isolated exercises, learners are immersed in a dynamic, project- centred curriculum. The learning journey is shaped by real-world projects that require active enquiry, critical analysis and collaboration. Throughout, AI-powered tools deliver continuous, tailored feedback by observing students’ approaches, prompting reflection and guiding them to deepen their reasoning. For example, a learner might analyse a complex dataset and produce a report, which can then be shared as a tangible demonstration of newly acquired skills. This method ensures that learning is interactive, iterative and directly tied to the practical demands of the workplace. AI at Work: From Productivity Hacks to Organizational Transformation 14

AI at Work from Productivity Hacks to Organizational Transformation 2026