When intelligence becomes abundant, judgement becomes the differentiator

I sometimes wonder whether I would have made it through the recruitment process that started my CA journey if an automated screening system had stood between me and that first conversation.

I stumbled into a room that PwC was recruiting for, on campus on my way to the volleyball court, without a completed application, without the forms and without a suit. A second year student, all geared up for the volleyball match, not a likely potential audit fit. I was curious. I asked questions. And I met someone willing to interview me and take a chance on me.

Would a system have allowed that conversation to happen? I cannot know. But a process requiring every box to be ticked before anyone spoke to me could easily have ended my journey there.

Someone exercised judgement while the information about me was incomplete. They gave themselves the opportunity to find out more.

I think about that now as I consider the workforce we are preparing to lead – and my own place in it.

Beyond knowing the answers

We have led people alongside machines for generations. What I am grappling with is the prospect of leading a workforce in which people and AI agents both carry out work we have delegated to them. People will use AI to extend their capabilities; agents will be given tasks and authority to act within agreed boundaries. As robotics develops, that relationship may become increasingly physical, too.

How do we lead well when the work is shared in this way? How do we decide whose recommendation to trust, what to delegate and when to intervene?

And, more personally, what will make my judgement worth trusting when an AI agent may produce a stronger analysis than I can?

I do not believe I, or anyone else, has all the answers. But I want to think about these questions now, for myself and for the leaders we are preparing for this future.

For much of our careers, we have been rewarded for knowing. Knowing the numbers, understanding the business, having the answer when someone turns to us in a meeting. As CFOs, we have worked hard to earn that confidence. It can become part of how we understand our own value.

When more of that analytical capability becomes readily available, we have to examine what we contribute once the information is on the table.

Knowing what matters

For me, judgement means deciding what matters, testing what we think we know and choosing a course of action while recognising what remains uncertain. It means being able to explain the choice and take responsibility for it.

There is something quite exposing about that. More intelligence does not relieve us of the need to make a call.

We are already surrounded by more material than we can meaningfully absorb. Reports arrive, dashboards refresh and another clip tells us what we should be thinking about. It is easy to keep consuming, hoping the next piece of information will make the decision obvious.

Before commissioning another analysis, can we describe the decision we are trying to make? What information could actually change it? What can we set aside?

Consider a pricing decision. If we ask an agent only how to increase revenue, we have already narrowed the answer. We also need to understand which customers can absorb a change, how they may respond and what it could do to retention. An attractive uplift may look very different once we consider the cost of replacing customers who leave.

Now imagine that the agent recommends a price increase, while the person closest to those customers raises concerns. I would want to understand what each view rests on. Is the concern supported by recent customer conversations? Has the analysis accounted for changing behaviour? Are we protecting a familiar approach because changing it feels uncomfortable?

The judgement begins with the objective we set and continues through how we examine the recommendation. Delegating the analysis still leaves us responsible for deciding whether we have understood enough to act.

This is also why I keep coming back to how we listen.

In a meeting, are we watching the person speaking, or looking down at our devices? Do we notice the hesitation before someone agrees? The raised eyebrow when a delivery date is mentioned? The person who has stopped contributing?

Sometimes the words say yes, but something makes us pause. Perhaps the team does not believe the timetable. Perhaps someone knows about a constraint that has not reached the executive discussion.

I cannot assume that an expression tells me what someone is thinking. My interpretation may be wrong. But I can notice, put my device down and ask: “What are we missing here?”

That moment of attention can bring information into a decision that would otherwise have remained unspoken. We should be careful about losing it as more of our work passes through systems and summaries.

The same applies when we assess people. One candidate may give every expected answer. Another may stumble over an answer but ask a question that reveals real thought, curiosity and a willingness to learn.

I would like to understand both more deeply. Enthusiasm matters to me, but confidence alone is an unreliable measure of capability. A quieter person may have a great deal to offer. What interests me is how someone reasons through an unfamiliar problem and responds when new information challenges their first answer.

My PwC experience reminds me how much depends on whether we leave room for that inquiry. I would like us to use technology to widen the opportunity to discover potential, while examining where our criteria might exclude it.

Challenging our own assumptions

Yet, I have to apply the same scrutiny to my own instincts. Being human does not automatically make my judgement sound. Experience can help me recognise something important. It can also make me too quick to believe I have seen it all before.

The physicist from Caltech, Richard Feynman put it plainly: “The first principle is that you must not fool yourself, and you are the easiest person to fool.”

I find that a useful challenge. How closely do I examine an analysis when it supports the answer I already favour? Would I apply the same scrutiny if it pointed in the opposite direction?

If I feel strongly about a decision, I should be able to explain what sits behind that feeling. Which assumptions am I making? What evidence would change my mind? Have I properly heard the person who sees it differently?

I would want to use AI to help challenge my reasoning and explore alternatives, while checking the evidence behind its responses. A persuasive explanation deserves examination, whoever or whatever produced it.

Delegating without surrendering accountability

Leading a mixed workforce will also require judgement about how much authority we delegate. A routine, reversible action may need a different level of oversight from a decision that materially affects a customer, an employee or the business. We will need to decide where people must remain involved and what should trigger a review.

Those boundaries deserve thought before something goes wrong. If an agent acts within instructions we have approved and causes an outcome we did not intend, we need to examine the instructions and our oversight. Saying “the system decided” would leave the leadership question unanswered.

I also want younger colleagues to see more of the reasoning behind decisions. If they hear only the final answer, they miss how we reached it. Let them hear why an assumption concerned us, why we sought another perspective and why we eventually felt we knew enough to proceed.

We should return to those decisions together. What did we expect? What happened? What would we now do differently? Recording our reasoning at the time gives us something honest to learn from, before hindsight makes everything seem more obvious than it was.

Making judgement worthy of trust

As I think about leading people and AI agents, I keep returning to these habits. I want us to remain curious enough to investigate what does not quite make sense, and willing to revise a view we have held with confidence.

Someone once exercised judgement that opened a door for me. As I prepare to lead people alongside AI agents, I think about the responsibility that comes with making decisions that shape another person’s opportunities, a team’s direction or a business’s future. I expect to be challenged by intelligence that exceeds my own in areas where I once felt most confident. I will need to know when to trust it. When to question it and when to change my mind. Whatever we delegate, we must still be able to stand behind the decisions made under our leadership.

When intelligence becomes abundant, our judgement must be worthy of the trust others place on us.