PwC analysed more than one billion job ads and found a two-track labour market: roles professionalised by AI, where judgement is emphasised, and roles democratised by AI, where the work gets easier for non-experts. Here is how to tell which track your role is on, read through Ability, Character and Environment.
Most of the argument about AI and jobs is asking the wrong question. It is not sorting people into safe and unsafe. It is sorting roles into two different tracks, and the same technology is doing both things at once.
In some roles AI strips out the routine work and leaves more of the judgement, so the person doing the job becomes harder to replace. In others AI makes the work itself easy enough that someone without the expertise can do it, so the person doing the job becomes easier to replace. Knowing which of those is happening to your role is the single most useful piece of information you can have right now, and it is knowable.
PwC's 2026 Global AI Jobs Barometer, published on 15 June 2026, put numbers on the split. Analysing more than one billion job advertisements across 27 countries and territories, PwC found what it calls a "two-track" labour market. Roles being professionalised by AI, where the technology automates routine tasks so that human judgement and expertise are emphasised, are seeing twice the growth in available jobs and 42% faster salary growth than roles being democratised by AI, where the technology makes the role itself easier for non-experts to perform. PwC's own examples of the first group include radiologists and recruiters. Its examples of the second include IT service managers and medical secretaries.
That is a structural finding, not a moral one. Nobody chose to be on the democratised track, and nobody earned their way onto the professionalised one. The tracks are a property of how AI happens to interact with the tasks in a role, and the same job title can sit on different tracks in different organisations depending on what the organisation decides to do with the time AI frees up.
Being professionalised means AI is taking the parts of your job you could do half-asleep and leaving the parts that need you awake. The work does not get lighter. It gets denser. More decisions per hour, less padding between them, and a higher cost when you get one wrong. People on this track often describe the change as pressure rather than relief, and that reading is accurate. The reason the market rewards it is that the remaining work is harder to hand to someone else, not that it is more pleasant to do.
The entry level is where this shows up most sharply. PwC analysed 2.4 million entry-level jobs in the US and found that early-career roles most exposed to AI are now seven times more likely to require traditionally senior-level human-intensive skills such as leadership, creativity, or face-to-face interaction. Openings for those roles have grown 35% since 2019, while other entry-level roles have shrunk by 10%. Pete Brown, PwC's Global Workforce Leader, described the mechanism directly in the release: "AI is removing some of the routine work that once acted as an apprenticeship, while increasing demand for judgement, leadership and adaptability much earlier in careers."
Read that carefully, because it cuts both ways. The junior roles that survive are better paid and more interesting. They are also being asked for capabilities that used to take a decade to build, without the decade.
Being democratised means the expertise that used to sit in your head now sits in the tool, and the tool is available to people who do not have your background. This is the harder side to sit on, and it deserves to be described honestly rather than softened. If the specialist knowledge that made you the person to ask is now a prompt away, the market value of that knowledge falls, and no amount of working harder at it will hold the line.
What does not fall is the value of everything the tool cannot do: knowing which question to ask, knowing when the confident answer is wrong, holding the relationship with the person who has to live with the decision. The democratised track is not a sentence. It is a signal that the centre of gravity in your role has moved, and that continuing to invest in the old centre is the actual risk.
It is also worth being precise about what PwC did and did not measure. The Barometer analyses job advertisements, company financial data and occupational task data. It tells you what employers are asking for and paying for at scale. It does not follow named individuals through their careers, so it cannot tell you what happened to any particular person on either track, and it should not be read as a prediction about you.
Four questions get you most of the way, and they are more reliable than a job title.
First, when AI does part of your work, what is left? If what remains is judgement, relationship, and accountability for the outcome, you are being professionalised. If what remains is checking the output and passing it on, you are being democratised.
Second, could a capable person outside your field now produce a passable version of your core deliverable with a good tool and an afternoon? If the honest answer is yes, that deliverable is no longer the thing you are paid for, whatever your contract says.
Third, is your organisation reinvesting the freed-up time or harvesting it? Two organisations can take the same AI capability and put the same role on opposite tracks. One uses the saved hours to push people into higher-value work. The other treats the saved hours as headcount. That is a decision made above you, and it is a fact about your situation rather than about your ability.
Fourth, what are the job ads for your role asking for now compared with three years ago? This is the cheapest research available and almost nobody does it. If the requirements have drifted toward judgement, stakeholder work and leadership, the track has already moved.
Adaptability Intelligence (AQ) is the measurable capacity to adapt, and AQai measures it through the A.C.E. model: Ability, what you can do; Character, who you are and how you naturally respond; Environment, the conditions you are adapting in. The two-track split lands differently on each of the three, which is why generic advice about "learning AI" so often fails to help.
Ability covers the developable skills of adapting, including Grit, Mental Flexibility, Mindset, Resilience and Unlearn. The instinct when a role starts being democratised is to reach for Grit, work harder at the thing you are good at, and out-run the tool. That is the wrong lever, and it is expensive. The lever the split actually calls for is Unlearn: the willingness to let go of a skill, a process, or a professional identity that has stopped paying, before it fully stops paying.
Unlearn is the least comfortable of the abilities because it asks you to give up the thing that made you competent, and competence is not a small thing to hand back. It is also the one the two-track market rewards most directly, because moving from the democratised side of a role toward the judgement layer of it is, by definition, an act of unlearning. Mental Flexibility does the adjacent work: holding more than one version of your future role in mind at once, so you are not betting everything on the version that flatters you.
Character covers your more stable patterns, including emotional range, thinking style, motivation style, and hope. Hope in the AQ sense is not optimism about the world. It is the belief that there is a route from where you are to where you want to be, and that you have some agency over taking it. It is the difference between reading the two-track finding as information and reading it as a verdict.
This matters because the behaviour that follows is so different. People with a route in mind try things, ask for the stretch project, and put their hand up for the work that is uncomfortable. People without one wait, hedge, and keep their heads down, which is a rational response to feeling that nothing they do will change the outcome, and which reliably produces the outcome they feared. Motivation style shapes how that route needs to look: some people need the destination to be vivid before they move, others need only the next step to be safe enough. Neither is better, but a development plan built for the wrong one will be quietly ignored.
Environment covers the conditions you are adapting in, including work stress, team support and company support. This is where most of the two-track story is actually written, and it is the part career advice tends to skip because it is not something the individual controls.
Whether the hours AI gives back get reinvested in you or taken off the payroll is an organisational decision. Whether you get access to the tools, the training and the higher-value work is an organisational decision. PwC's own workforce leader concluded that organisations need to rethink how they develop talent if they want people to thrive in this environment, which is a plain statement that the responsibility does not sit with the individual alone. When someone is not moving up the track, the first question worth asking is whether they lack the ability or whether they have never been given the conditions. Treating an Environment constraint as a personal failing is the most common and most damaging misdiagnosis in this whole conversation, and it wastes people who could have adapted perfectly well in a different setting.
If you are on the professionalised track, protect the judgement. The risk on this side is not redundancy, it is depletion: denser decisions, thinner recovery, and a slow erosion of the quality of thinking that the role now depends on. Guarding capacity is a performance decision, not a wellbeing perk.
If you are on the democratised track, move early and move up rather than sideways. Sideways means finding another task the tool has not reached yet, which buys you a shorter runway than it feels like. Up means moving toward the parts of the work that involve deciding, judging, and being accountable, and being willing to be a beginner there for a while. Ask for one piece of work above your current level and let it be visibly uncomfortable.
If you lead people, stop asking whether your team is adaptable and start asking what your team is being asked to adapt to, and whether the conditions you have built make that possible. Name which roles in your team are being professionalised and which are being democratised, out loud, and say what you intend to do about the second group. People can handle a hard truth about their role far better than they can handle a vague reassurance that everything will be fine. The organisations that keep their people through this split will be the ones that told them the truth early enough for the truth to be useful.
Measuring adaptability makes that conversation specific instead of speculative. An AQme assessment shows whether the block is an Ability the person can develop, a Character pattern that shapes how they will approach it, or an Environment constraint that no amount of individual effort will fix, which is the difference between a development plan that works and one that quietly blames the wrong thing. You can see how the three dimensions fit together in the AQ model, and the entry-level side of this split is covered in more depth in our piece on how AI is erasing the first rung of the career ladder.
Both, in different roles at the same time. PwC's 2026 Global AI Jobs Barometer, published 15 June 2026, describes a two-track labour market: roles it calls professionalised, where AI removes routine work and human judgement is emphasised, are growing at twice the rate and with 42% faster salary growth than roles it calls democratised, where AI makes the job easier for non-experts to do. The technology is the same; the effect depends on what is left in the role once the routine is gone.
Look at what remains after AI does a piece of your work. If what is left is judgement, relationships and accountability for the outcome, the role is being upgraded. If what is left is checking and forwarding the output, it is being commoditised. A second check is whether a capable outsider could now produce a passable version of your core deliverable with a good tool and an afternoon.
PwC found the average wage premium for workers with AI skills reached 62% in its 2026 Barometer, up from 57% the previous year, and that jobs requiring specific AI skills such as prompt engineering or machine learning grew 69% against 9% for the overall jobs market. The premium varies widely by sector, from as high as 118% in consumer markets to 16% in government and public sector work, so the headline number is an average rather than a promise.
Adaptability Intelligence (AQ) is the measurable capacity to adapt, and AQai measures it across three dimensions: Ability, the developable skills of adapting such as Unlearn and Mental Flexibility; Character, the more stable patterns such as hope and motivation style; and Environment, the conditions you are adapting in. Specific AI skills have a short half-life because the tools change. The capacity to keep letting go of what has stopped working, and to keep learning what has started to matter, does not.
It is shared, and the organisational share is larger than most career advice admits. Whether the time AI frees up gets reinvested in people or taken as savings, and whether individuals get access to the tools, the training and the higher-value work, are decisions made at organisational level. PwC's Global Workforce Leader, Pete Brown, concluded that organisations need to rethink how they develop talent if they want people to thrive. In AQ terms, an Environment constraint should not be diagnosed as an Ability deficit.
Copyright adaptai Ltd. All rights reserved. AQ® and AQai® are registered trademarks of adaptai Ltd.
.png){kind=small}
.png)
