People come first, and AI belongs in second place
The phrase has hardened into orthodoxy. AI with humans in the loop. It appears in strategy decks, vendor pitches and government white papers, and it sounds reassuring, as though the human has been safely accounted for. André Lacroix thinks it has the whole thing backwards. In his telling, putting AI at the centre and humans at the edge is precisely the error leaders are about to make at scale. The future he describes is the reverse: purpose-based, people-centric leadership with AI in the loop. The words are almost identical. The order of priority is not.
André is Chief Executive of Intertek, a FTSE 100 quality assurance business with more than 44,000 colleagues across 100 countries. He is not a technologist selling a platform, and he is not a sceptic dismissing the technology. He is a global operator deploying AI across a regulated, safety-critical enterprise, which is what makes his refusal to lead with the technology worth taking seriously.
Your AI advantage will not survive contact with the market
The strategic core of André's argument is a prediction that many boards will not want to hear. Every algorithm and every agentic solution, meaning AI that can act and complete tasks with limited oversight, will eventually become a commodity. The tool that feels like a competitive edge this year becomes standard issue the next. If that is true, then the millions currently being spent to acquire a technology advantage are buying something with a short shelf life.
What remains once the technology is equal across competitors is people. André calls this EQ supremacy, and he is building his next book around it, a sequel to Leadership with Soul. His claim is that emotional intelligence, the capacity to engage, energise and lead people, becomes the last durable differentiator in an economy where everyone has access to the same models. I find this the sharpest strategic point in the conversation, because it reverses the usual investment logic. The spending that looks like a cost, developing people, turns out to be the part competitors cannot copy.
Link to Maggie Sarfo Episode 16
The CEO avatar is the wrong answer to the engagement crisis
André grounds his people-first argument in a number. Gallup's latest workplace survey shows 80% of the global workforce is not engaged or not actively engaged. He translates that into three billion people arriving at work each day without purpose, and he calls the corporate world "over managed and underled." His case for AI is that it can automate the analytical work filling a leader's diary and hand that time back to be spent on people.
Which is exactly why he rejects the CEO avatar. Asked about leaders building digital twins of themselves to answer employee questions at scale, his response was unusually blunt. A chief executive, he argues, is a chief energy officer, and the entire point of the role is direct human engagement. Outsourcing that to a bot does not free the leader; it removes them from the one task the job exists for.
The counterargument sits inside his own position, and to his credit he does not hide it. He wants "data science in real time at scale," processing three million data points in his own role and half a million for a regional leader. Volumes that large cannot be held in any human head. So the honest question the episode raises, and does not fully settle, is this: as the machine handles what people cannot, how does human judgment stay genuinely in the lead rather than nominally in charge?
Instinct is data, and leaders discard it at their cost
André's answer leans on a reframing of instinct. He tells students not to fear their gut, because experience is years of data processing accumulated in the brain. A seasoned leader's instinct, on this view, is not the opposite of analytical rigour but another data point produced by it. He advises leaders to size the opportunity quantitatively, using financial and non-financial metrics, to understand how every stakeholder feels, and then to step back and let both the analysis and the instinct speak.
That is a more demanding standard than it first appears. It refuses the easy move of letting the apparently objective numbers override a leader's judgment, and it refuses the lazy move of calling gut feel a substitute for doing the work. The consequence for leaders is that neither the algorithm nor the instinct gets the final word alone.
Responsible AI is four disciplines, not a press release
The most immediately usable part of the conversation is André's framework for adopting AI responsibly, drawn from Intertek's own experimentation. He names four foundations. Governance, meaning a committee and processes that understand both the opportunities and the risks. Transparency, meaning technical documentation precise enough that an algorithm can be audited at any time. Security, meaning cyber protocols fit for an unproven technology. And functional performance, meaning accuracy standards set high enough to trust.
He is scathing about that last one. Providers have pitched him tools at 40, 50 and 60% accuracy, and he argues that shipping AI at those levels is how organisations teach their own people that AI is worse than a human. He is equally direct about a risk many firms ignore: uploading company data into a third-party large language model, one of the AI systems behind common chatbots, can mean that data leaves the company altogether. He offers no comfort on this and I will not add any that he did not; his point is that many firms are exposed and do not know it.
Underneath all four foundations is a claim about who should set the pace. Regulators, he says, are behind the curve in every industry Intertek works in, and he does not blame them, because the task is genuinely hard. The best companies, in his view, do not wait. They set standards well beyond the regulatory minimum, and he argues AI demands exactly that posture.
The challenge André leaves for leaders
The uncomfortable takeaway is that AI does not let leaders off the hook; it raises the bar on the human parts of the job. If the technology commoditises, the differentiator is people. If instinct is data, judgment cannot be delegated. If regulators lag, the standard is yours to set. André closes on a line he gives students, and it applies just as well to the executives listening: if you want to change the world, you have to change yourself. The reinvention AI demands starts with the leader, not the tool. The question worth sitting with is whether your organisation is spending its AI budget on the thing that lasts, or the thing that will be copied by this time next year.
Listen to the full conversation with André Lacroix on The AI Adoption Podcast:
Comments