Five years ago, my interview questions were simple. Could the candidate build the model? Pull the cash position? Close the books cleanly and on time? Speed and accuracy were the currency, and the person who produced the right answer fastest moved ahead.
These days, a weekly cash reconciliation that once consumed an entire working week now arrives in ten to fifteen minutes. The system can flag anomalies, model scenarios, and generate a board-ready summary. No one needs to stay late searching for a missing £4,000.
When systems generate the number faster and more reliably than a human can, speed becomes largely irrelevant. And the profession is dividing. On one side: transactional, process-heavy work that AI absorbs quickly. On the other: growing demand for judgment-driven, communication-focused skills. The Big Four are already feeling this pressure, with KPMG cutting around 600 UK roles this year.
But a 2026 National Bureau of Economic Research survey of more than 750 CFOs tells the fuller story. AI drove a 1.8% increase in labour productivity last year; demand for data scientists, engineers, and analysts is projected to grow by 1.35% by 2028. And 68% of CFOs identify skills gaps, not technology, as the main barrier to capturing value from AI.
So what does it take to get hired in finance today? Here are the five skills that will get you noticed.
-
AI Literacy and Agent Orchestration
The baseline expectation has shifted. Knowing how to use ChatGPT is not a skill. The skill is knowing how to prompt, chain, monitor, and correct AI agents for invoice matching, reconciliation, anomaly detection, and dynamic scenario planning.
Many CFOs who now act as strategy leaders have fully integrated AI agents into their finance functions. Junior professionals with relevant experience are being brought into strategic conversations earlier than any previous generation.
The professionals who thrive will be the ones who realised the Excel spreadsheet was never the point. The number was always the means to a decision.
-
Advanced Data Analysis and Predictive Modelling
Move beyond reporting, the 2026 requirement is building scenario models that drive revenue decisions; using Python or R basics, machine-learning-assisted forecasting, and total-visibility auditing rather than sampling.
Deloitte identifies data analysis and AI fluency as the top skills priority for finance leaders in 2025–2026. FP&A now demands predictive modelling as a core function, not a specialisation.
If you are still building your professional identity around monthly reporting and Excel, you are building on a foundation that is being automated away beneath you.
-
Critical Thinking and the Ability to Spot What AI Misses
This is the skill I now listen for most carefully in interviews – the moment a candidate describes catching something the tool missed.
A close friend working in one of the world’s top scale ups received an AI-generated board summary. The chart was clean and the formatting precise. However, the underlying assumption was wrong: an FX rate applied to the wrong entity, compounding through every figure that followed.
No one flagged it before the meeting. No one questioned it during, either, because it looked authoritative. The professionals who add real value do three things consistently: they ask what assumption produced the number; they trace where the data originated; and they question whether the model has seen a situation like this before.
-
Strategic Narrative and Business Storytelling
If the system produces the output, what remains are conversations – the board update, the CEO briefing, the explanation of why the cash position looks the way it does and what it means for next quarter. The CFOs who do this well are not the ones who recite figures. They are the ones who change what happens next. They can make a £2m variance feel tangible to someone who has never opened a P&L.
That skill was always valuable, but it is more exposed now. The grunt work that once filled the week, and often masked weaker communicators, is being absorbed elsewhere.
The likes of McKinsey, Boston Consulting Group, and Bain are using AI to speed up research, generate insights, and produce client materials in a fraction of the time. The human who remains essential is the one who can shape what the narrative means and what action it demands. High-performing finance teams are 2.3 times more likely to cite emotional and social intelligence as a success factor.
-
Adaptability and the Builder’s Mindset
The strongest candidate I interviewed recently did not stand out because of qualifications. Outside of work, they had built a small AI application (unrelated to treasury or FP&A) simply to understand the tools. A friend at J.P. Morgan created an app to track kite-surfing conditions. A colleague built a daily habits tracker. Neither project was financial. Both revealed the same highly desirable profile.
Each required identifying a problem, choosing a tool, hitting friction, improvising, and shipping something functional. A training certificate proves attendance. It does not prove initiative once the instructions end.
The question I ask every candidate now is: What have you built or explored using AI in the last six months, outside of work? It is not about finance or polish – it is about motion. When there was no deadline, no manager, and no requirement, the answer tells me more than anything on a CV.
The Cost of Waiting
The cost of waiting is asymmetric. Skills compound slowly, so developing them after reconciliation work has already disappeared leaves you at a disadvantage.
You are entering at a moment when automation is stripping away low-value repetition faster than any previous generation experienced. That forces you upward – toward validation, strategy, and decision-making – earlier than your predecessors. That is a far more interesting career than matching transactions ever was.
