For decades, decisions about who to hire, promote and reward were guided largely by instinct. Leaders prided themselves on their ability to “read people”, trusting experience over evidence. But as organisations become more complex and stakes rise, that approach is being challenged. In the latest episode of Think Ahead, Sergei Guriev, Professor of Economics and Dean of London Business School, spoke with Isabel Fernandez-Mateo and Khyati Sundaram about how data and artificial intelligence are transforming people management, and why human judgement still matters.
Isabel Fernandez-Mateo, Adecco Professor of Strategy and Entrepreneurship at London Business School, has spent more than two decades studying how organisations make decisions about people and how those choices shape careers and inequality. Khyati Sundaram, former CEO and founder of Applied, brings a practitioner’s perspective, having built a data-driven hiring platform designed to reduce bias and improve real-world outcomes.
Why people decisions are changing
The shift towards evidence-based people management has been driven by three forces. First, organisations now collect far more data than they once did, from applications and performance metrics to digital traces of collaboration. Second, analytical tools, including machine learning, have become more powerful and accessible. Third, decades of research have shown that intuition is often flawed.
As Isabel explains, human judgement is vulnerable to cognitive biases, from first impressions to favouring people who feel familiar. These instincts can feel like expertise, but they are not always predictive of performance. Yet letting go of intuition is hard. For many managers, relying on experience feels like a core part of leadership.
Hiring in the age of AI
Hiring is where this tension is most visible. Despite strong evidence that structured interviews outperform informal conversations, many organisations still rely on unstructured “chats”. These interviews offer a sense of chemistry and control but are not always about job-relevant ability.
Khyati has seen this resistance first-hand. Even when organisations adopt analytics, they often use data to confirm what they already believe. Confirmation bias can turn analytics into a veneer of objectivity rather than a tool for better decisions.
Generative AI has added a new layer of complexity. Employers are using AI to define roles, target candidates and screen applications. At the same time, candidates are using AI to write CVs and tailor applications at scale. The result, as Khyati describes it, is a “sea of sameness”, where applications look increasingly alike, making it harder to distinguish genuine capability.
In response, some organisations are reverting to older signals such as networks or educational pedigree because they are harder to replicate with AI. Others are adding more layers to the hiring process, from work samples to additional assessments, to test authenticity. While these approaches may help, they also increase cost and complexity.
Performance, fairness and unintended consequences
Beyond hiring, measuring performance remains one of the hardest challenges in people management. More data does not necessarily mean better measurement. Objective metrics can be gamed once they become targets, while subjective evaluations are vulnerable to bias.
Isabel argues that no single measure is sufficient. Performance is shaped by context, teams and circumstances, not just individual effort. The goal should be a thoughtful combination of objective and subjective measures, interpreted with care. Data can support these judgements, but it cannot eliminate uncertainty.
Why judgement still matters
Both guests agree that AI excels at spotting patterns at scale, but it cannot fully account for context, trade-offs or organisational priorities. Leaders still need to interpret insights, make difficult choices and explain decisions that affect real people.
Isabel emphasises that analytics should increase confidence in decisions, not make them on leaders’ behalf. Khyati adds a cautionary note against what she calls “algorithmic confidence laundering”, where leaders hide behind data to justify decisions they have already made. Instead, data should prompt better conversations and more accountable judgement.
What leaders should do now
The takeaway is clear. Using data and AI in people management and hiring is not just a technical exercise. It is a leadership challenge. Organisations need to invest not only in tools, but also in capabilities, culture and governance. Asking better questions, understanding what data can and cannot tell you, and remaining accountable for decisions – these are all more important than ever.
In the move from intuition to evidence, the goal is not to remove humans from the process. It is to help leaders make better, fairer and more thoughtful decisions about people.