Last December, during my final year of studying Economics at Leeds, I secured a graduate position at a boutique, London-based strategy consultancy. In an increasingly difficult market, I felt proud of the hard work and planning that had gone into securing a high-quality graduate role; the culmination of 23 years of education, internships, and work experience. I felt fortunate to have avoided the prospect of unemployment or underemployment.
However, five months later, during my final exams, a five-line email arrived in my inbox explaining that my offer had been rescinded due to “rapidly evolving changes in our business needs”.
Naturally, I immediately picked up the phone and dialled the company to find out what had happened. A short call with one of the firm’s partners translated the phrase into plainer English: I, along with the rest of that year’s graduate intake, had been replaced by AI.
Such a candid admission of their reasoning was unusual, but my general experience was not. Competition for graduate employment has reached record levels. According to the Institute of Student Employers, companies now receive an average of 140 applications for every graduate vacancy, compared with 38 two decades ago. ‘Graduate’ remains the lowest earning category on Adzuna below Domestic Help and Cleaning, a UK employment tracker that collates information from hundreds of job boards.
Meanwhile, the number of graduates entering the market continues to grow. According to UCAS, a record 262,820 UK 18-year-olds secured university places on results day this year. For those who struggle to make the transition into work, the effects can become self-reinforcing. The government-commissioned Milburn report found that prolonged unemployment can erode confidence, allow skills to deteriorate and make returning to work harder.
There are several possible causes at play, all interacting in unknowable proportions. Employers face weak economic growth and higher employment costs, while remote work may be making experienced hires more attractive than graduates who require closer supervision. AI is also changing both the work available to junior employees and the way candidates apply for it.
There are competing views about where this leads. Optimists expect AI to raise productivity, create new forms of work and free people from repetitive tasks. Sceptics fear continued job displacement, unproven business models and investments into hyperscalers that run ahead of sustainable demand. At this stage, certainty in either direction still seems premature.
Nevertheless, my own story offered a striking enough-example of AI affecting white-collar employment directly that a reporter contacted me about it, leading to coverage in The Sunday Times and interviews with Times Radio and BBC Radio Leeds.
Then, the following unexpected LinkedIn message arrived:
“Nice article. We are looking for a grad and you may find what we are doing with AI to enhance, not replace, our great people really interesting.”
The message was from Charlie Whelan at Polestar Corporate Finance.
Our subsequent conversations made clear that Polestar did not see investment in people and investment in technology as a zero-sum choice. The firm was prepared to invest in my professional development while giving me access to tools that could accelerate it. We agreed that AI is a fantastic tool for enhancing efficiency when prompted effectively and deployed in moderation but should not be used to outsource critical thinking skills or produce un-checked outputs.
Since joining Polestar, it has been clear that this ethos is more than hot air. Its use of open-weight models out-does all AI deployments I’ve encountered in similar companies even though all outputs are still challenged, checked and placed in context by people. Used well, AI can accelerate research, interrogate information and remove some of the repetitive, predictable work involved in analysing a business. Used recklessly, it can lead to embarrassing mistakes.
That matters particularly in corporate finance. A model can process information quickly, but it cannot assume responsibility for the advice given to a client. Nor should efficiency come at the expense of the judgement, trust and critical thinking on which that crucial advice depends.
There is also a longer-term question for employers. Junior colleagues require training and initially consume some of the time of their more experienced colleagues. They also become the experienced employees of the future. Businesses that automate away their entry-level roles may reduce costs today while weakening their future talent pipeline. On a large enough scale, what might appear to be a rational cost-based exercise today may create a collective action problem: fewer opportunities for young people to develop, fewer experienced professionals for business to recruit later, and eventually mass unemployment.
This does not mean protecting every task from automation. It means distinguishing between tasks that require judgement and tasks that merely consume time. The aim should be to automate the latter so that people have more capacity for the former.
Losing my original graduate role showed me how disruptive AI can be. Joining Polestar has shown me another side of the same technology: used with care, it can help people learn faster and contribute more.
I am excited to be starting my career in corporate finance with a motivated, friendly team that is embracing AI without overlooking the value of its carbon-based employees.
