Opinion: Pearson CTO on the answer to the AI productivity gap

As AI is integrated into the workplace, everyone is asking the question of what AI will do to jobs. While headlines on rising unemployment fuel debate about the future impact, there is a more immediate question: if AI is everywhere, why aren’t workplaces seeing the productivity jump they were promised? If AI is spreading quickly while output per worker barely moves, something fundamental isn’t working.

Expanding access to AI tools matters, but access alone won’t close the productivity gap. Learning is what turns technology into results. That distinction is where much of the current conversation goes wrong.

The real constraint on productivity

What’s holding productivity back isn’t technology, but learning capacity. Learning is the missing link between AI promise and performance.

But learning is hard, and it’s supposed to be. It takes courage – courage from business leaders to integrate it effectively, and courage from individuals to adapt to new technologies. This demands a new approach to learning, supported by new solutions that more naturally support learners in the flow of work. Human judgment, AI assistance, and feedback need to reinforce one another continuously not through one off trainings, but inside daily operations. Productivity improves when work is redesigned at the task level and learning is built directly into how that work gets done.

Take a customer-support team. AI can draft a response in seconds, but the real productivity gain comes from the surrounding ecosystem, which requires a level of human input – setting up a verification or scheduling a lightweight review step for tougher cases. Over time the prompts, and policies improve, along with the human-led team. This is because the learning is happening naturally inside the workflow instead of a one-off training course.

Learning embedded in the flow of work

So, what steps can we take to ensure that in an AI-driven economy, learning itself becomes infrastructure?

First, learning must be embedded directly into the flow of work.

That means feedback while the work is happening via personalised coaching and AI guidance integrated into daily tasks that helps people choose their next best action. It also means teaching simple learning habits, like setting a goal for the week or trying a new approach, and noting what worked in order to adjust accordingly. When those habits are reinforced by the tools people already use, skill-building becomes faster and more durable.

The second shift is a change in approach, which starts at the top levels of an organisation. If learning is left as an afterthought, it becomes optional. It must be an integral part of the process right from the start and implemented from the top down. Business leaders need to work together to rethink how workflows can be redesigned, and champion the adoption of a common skills ontology.

But transformation cannot sit solely with the CTO, just as workforce strategy cannot sit solely with the CHRO. The most effective organisations are treating the CTO and CHRO as co‑architects of transformation. Together they can answer the questions that matter, such as where does AI remove bottlenecks and how does that free people to focus on higher-value work? With the right operating systems in place, AI shifts from ‘interesting experiments’ to ‘system-level performance improvements.’

Closing the productivity gap

Leaders must rebuild workflows around human judgement, intelligent systems and continuous skill development. Every positive scenario for an AI-enabled future is built on human development.

The winners will be the organisations who close the learning gap by embedding learning into work and treating human development as the foundation of productivity in the age of AI.