How AI is changing the role of economists
As an economist and CEO, I’ve been reflecting on what AI means for the work we do at Oxford Economics. Like many organisations, we have been experimenting extensively with AI, both in how we work internally and in how we deliver our research to clients. The experience has reinforced my belief that AI will not replace economists. It does, however, give us an opportunity to refocus where we add the greatest value.
AI is already changing how economists work
Much of the discussion around AI focuses on what it might one day achieve. I think it is more useful to start with what it can already do today. Like any economist, a significant part of my career has involved reviewing forecasts, checking consistency across datasets, validating assumptions and quality-assuring outputs. These tasks are essential to ensure our analysis is robust and trustworthy and to help economists develop the strong critical judgement that sits at the heart of our profession. But they are also time-consuming and, in many cases, highly repetitive.
This is where AI is already delivering tangible value. At Oxford Economics, we have been using AI to strengthen our quality-assurance processes. Traditionally, automated checks have relied on fixed programming, for example flagging forecasts that exceed a particular threshold. AI allows us to go much further. It can review thousands of forecasts simultaneously, identify unusual patterns, recognise inconsistencies that warrant further investigation and provide economists with another set of eyes before analysis reaches clients. It can also recognise recurring issues over time, helping us identify repeated patterns that conventional automated checks would struggle to detect.
That naturally raises a question: If AI can reliably perform many of these supporting tasks, where how should our economists be re-allocate their time?
To answer that, we need to think about what people have always valued from economists.
In a world where information becomes abundant, what becomes scarce?
Generative AI has made information dramatically easier to access. It can summarise lengthy reports, retrieve relevant research and generate answers in seconds. These are remarkable capabilities, and they will continue to improve.
But access to more information is not necessarily what organisations need most. They need to understand what matters, what might happen next and how they should respond. As information becomes more abundant, the insights that help answer those questions become more valuable.
These are exactly the kinds of questions we have spent decades helping clients answer at Oxford Economics. We have built a large body of rigorous economic analysis to help organisations make sense of developments in the global economy and understand what they mean for their business.
AI now gives us new ways to deliver that expertise, making it more accessible, more personalised and more deeply embedded in our clients’ decision-making. New tools we built such as AskOEAI and our integration with the Model Context Protocol (MCP) are helping our clients access our insights more naturally within their existing AI tools and workflows.
But the quality of what those tools deliver still depends on the expertise behind them. Behind every AskOEAI answer sits years of research, along with months of testing and training by our economists to ensure the outputs are accurate, grounded in evidence and worthy of our clients’ trust.
And that, for me, points to the real opportunity AI creates for economists.
The economist’s job is evolving, and that should excite us
If AI can support more of the repetitive work around economic analysis, while also giving people better ways to access our expertise, we have an opportunity to spend more of our time strengthening the thing that ultimately creates that value: the economics itself.
Good economics has never been about simply retrieving or restating what is already known. It is about exercising judgement, challenging assumptions, interpreting conflicting evidence and identifying signals that others have not yet recognised.
None of that is new. In fact, it is what we have always expected from great economists. What AI changes is the opportunity, and perhaps the urgency, to focus more of our time and development on those capabilities.
That means continuing to develop economists and analysts who can exercise sound judgement, produce rigorous analysis and communicate their conclusions clearly and honestly. Our approach to responsible AI adoption should help us create more space for those capabilities, not diminish the value of human expertise. I see that as a welcome opportunity.
Alongside applying AI within our own work, at Oxford Economics we are developing new ways for clients to access and apply our research, data and forecasts through our AI and data solutions.
