You May Be Asking the Wrong Question About AI

The most common question I hear from leaders about AI is some version of: ‘Is my job going to be safe?’

It’s understandable. But it’s the wrong question. Asking it keeps people in a posture of anxiety rather than inquiry.

A more productive starting point is: which parts of my job are already being replaced, and what does that mean for how I’m spending my time?

Most organisations don’t know which of their skills are already obsolete. Not which roles are at risk - which specific skills, inside specific roles, in their organisation. Generic research tells you that 40% of customer service jobs face automation risk. That’s not really actionable in any way. What actually helps in making decisions is knowing whether it’s the inquiry-handling or the relationship management that’s going - and what skills your people should be building instead.

TalentJam’s whitepaper, Navigating the AI Skills Transition, introduces a three-category framework that gets to that level of specificity. Skills facing full replacement are routine, rule-based, and high-volume - tasks AI can execute faster, cheaper, and at scale. Skills facing partial augmentation still require human judgment, but AI compresses the time and effort required significantly. Evergreen skills - leadership, conviction under ambiguity, relationship complexity, ethical judgment - remain genuinely resistant to displacement.

The framework produces some counterintuitive findings when applied to roles people assume are protected. Take investment research. The obvious target is the number-crunching - remembering that structured data analysis has been partially automated for years. What closer analysis actually flags as higher risk is the written output: first-pass synthesis, sector summaries, standard investment memos. Competent, templated, replicable. The genuinely evergreen skill is conviction formation under ambiguity - recognising when the model is wrong, reading what the data isn’t capturing, making a call that can’t be fully explained by the inputs. That has real implications for who you hire, who you develop, and what you ask senior people to spend their time on.

The pace of change makes this more urgent than most organisations are treating it. AI-related job requirement growth ran at over 7,000% in customer service roles between 2018 and 2024. The share of HR leaders actively deploying generative AI jumped from 19% to 61% in just eighteen months. The horizon has shifted from decades to years - and in some cases, months.

Applied to SFIA 9 - a globally recognised taxonomy of 147 technology and digital skills - the framework produces a striking result: just 7 skills are replaceable today, 118 are partially augmentable, and 22 are genuinely evergreen. If your workforce has significant concentration in those 7, and in roles built around the 118, the capacity math changes materially. Whether you know where your concentration sits is a different question.

There are three things worth establishing clarity on. Can your organisation answer which specific skills face displacement - not by function or role family, but by skill? Are you still recruiting for capabilities AI will replicate within your planning horizon? And when AI handles the routine work, have you decided what your people will do instead - or are you assuming the answer will become obvious?

For most organisations, the honest answer to all three is no. Not because the problem is intractable, but because the analysis hasn’t yet been done at the right level of granularity. Broad reassurances about the irreplaceable value of human judgment sit alongside hiring decisions and development investments that haven’t yet caught up with what’s actually changing.

The leaders and organisations that navigate this well won’t be the ones that resolved their anxiety about whether their roles will be safe. They’ll be the ones who stopped asking whether AI was coming for their role, and started asking what skills they should develop to take them into the future.