Career Capital After Automation

A senior analyst built her reputation on flawless financial models. Then she watched a junior colleague produce a comparable model in twenty minutes with an AI copilot. The thing she was known for — the thing on her résumé — had quietly become a commodity. What her managing director now pulls her into the room for is the other thing she does almost unconsciously: knowing which assumptions to distrust, which client will balk at the third slide, and which number is technically correct but strategically wrong. The skill she spent a decade perfecting is depreciating. The judgment she never thought to name is the asset that is appreciating.

This is the quiet repricing happening underneath every knowledge-work career right now, and most professionals are not accounting for it. We tend to treat a career as a ladder — a sequence of rungs you climb by getting better at the thing you already do. A more useful model is a portfolio: a collection of assets — skills, reputation, judgment, relationships, a track record — each with its own value, and each subject to a market that decides what is scarce. Automation has not destroyed that portfolio. It has changed the prices. The strategic question is no longer “how do I get better at my job?” It is “which of my assets is the market about to reprice, and where should I reallocate?”

Career capital is a portfolio, and the regime just changed

The scale of the repricing is not subtle. The World Economic Forum projects that 39% of the average worker’s existing skill set will be transformed or rendered outdated between 2025 and 2030 [1]. LinkedIn’s economic research puts the figure higher still, expecting roughly 70% of the skills used in most jobs to change by 2030 [2]. Underneath the WEF’s headline net gain of 78 million jobs sits enormous churn — 170 million roles created and 92 million displaced [3]. A net number that mild conceals a turnover that violent.

What makes this a regime change rather than ordinary progress is which assets are being devalued. For most of the automation era, the machine came for routine, lower-wage, physical work first. Generative AI inverts that pattern. McKinsey estimates current AI could automate activities that absorb 60 to 70% of employees’ time [4], and — crucially — finds it has more impact on knowledge work in occupations with higher wages and educational requirements than on other kinds of work [5]. The capital most exposed to repricing is precisely the codifiable expertise that ambitious managers and senior specialists spent their careers accumulating. Standing still is not safe; it is a concentrated bet on an asset class in decline.

What is being commoditized: codifiable knowledge and rote execution

It helps to be specific about what is losing value, because the answer is not “technical skill” in the abstract. It is the codifiable portion of it — the part that can be written down as a procedure and therefore handed to a machine. The WEF’s own skills outlook shows the clearest declines in exactly these areas, with manual dexterity, endurance, and precision posting the largest net drop and 24% of employers expecting their importance to fall [6]. The same logic now reaches into cognitive work: anything that reduces to a repeatable transformation of inputs into outputs — the standard model, the first-draft memo, the routine analysis — is becoming abundant where it used to be scarce.

Abundance is the operative word. When a capability becomes abundant, its price falls, regardless of how hard it was to acquire. The decade you spent becoming excellent at a now-automatable task does not entitle you to a premium the market no longer pays. This is the uncomfortable part of the portfolio framing: sunk cost is not asset value. The skill churning fastest is concentrated where AI bites hardest — PwC finds the skills employers ask for are changing 66% faster in the occupations most exposed to AI [7]. In those roles, the half-life of what you know is collapsing.

The repricing in real time: what the market now pays for

If some assets are falling, others are rising sharply — and the wage data shows it happening in real time. PwC’s 2025 analysis found that the premium employers pay for AI skills more than doubled in a single year, reaching an average of 56% [8], while productivity growth nearly quadrupled in the industries most exposed to AI [9]. That headline premium is weighted toward specialized technical builders, but the upside is not confined to them: a separate Lightcast study put the salary premium for general AI-related skills at 28% — nearly $18,000 a year — and found it spreading well beyond the technology sector [10]. These are not the wages of people being replaced. They are the wages of people who have learned to direct the technology rather than compete with it.

That distinction — direct rather than compete — is the heart of the reallocation. The most telling evidence is how AI is actually used on the ground: Anthropic’s analysis of real usage found 57% of tasks augmenting human work versus 43% automating it [11]. The dominant mode is augmentation, which means the scarce, well-paid position is not “person who does the task” nor “person whom the task was taken from,” but “person who wields the tool with judgment.” The capital that appreciates is whatever makes you the second kind of person.

What is appreciating: judgment, taste, and verification

The first appreciating asset is judgment — the ability to decide what is worth doing, to spot the answer that is plausible but wrong, and to know where the tool can be trusted. The evidence here is sharper than the usual hand-waving about “human skills.” In the landmark BCG and Harvard study of consultants using generative AI, those with access to the tool produced work more than 40% higher in quality — but only on tasks within the technology’s capabilities [12]. Outside that “jagged frontier,” the tool confidently produced worse work, and the people who did best were the ones who knew which side of the frontier they were on. Knowing where the machine is reliable, and where its fluent output is quietly wrong, is itself a high-value skill.

This is why verification is becoming a core competency rather than a clerical one. As reliance on AI grows, knowledge workers are shifting from doing the task to overseeing it — a transition that researchers warn can erode critical engagement precisely when it is most needed [13]. The risk and the opportunity are the same fact: when generating a draft costs almost nothing, the value migrates to the person who can rigorously check it. The professional who treats AI output as a confident intern’s first draft — useful, fast, and never to be shipped unexamined — is allocating capital correctly. The one who treats it as an oracle is letting their most valuable asset, their judgment, atrophy.

The human skills that compound

The second appreciating asset class is the set of distinctly human capabilities that resist codification, and the demand data is now unambiguous. McKinsey projects demand for social and emotional skills rising 26% in the United States through 2030 [14]. An analysis of more than 75 million job postings found that eight of the ten most-requested skills are now durable human skills like leadership and communication — up from seven of the top ten a few years earlier [15]. Notably, even the WEF’s top “core” skill is not a technical one: analytical thinking remains the single most-valued capability, called essential by seven in ten employers [16].

The portfolio implication is not the tired advice to “work on your soft skills.” It is that these capabilities are appreciating because the technical ones are being commoditized — they are the complements that rise in value as their substitutes get cheap. When a competent first draft costs next to nothing, the binding constraint shifts to the things AI cannot do: aligning people who disagree, earning trust across a boundary, reading a room, framing a decision so an organization will actually act on it. These were always valuable. What has changed is that they are no longer the soft accompaniment to the “real” technical work. Increasingly, they are the work.

The experience paradox

Here the picture turns genuinely counterintuitive, and it is worth sitting with the tension rather than resolving it too neatly. Several studies find that AI’s productivity benefits flow disproportionately to the less experienced: in one widely cited field study, the least-skilled workers gained up to 35% while the most experienced gained little [17]. Read narrowly, that is a threat to the experience premium — if a tool lifts a novice to near-expert output on a given task, the market value of having spent years getting good at that task compresses.

But that is only true for the codifiable task, which is exactly the asset we have already established is being devalued. The deeper, less substitutable form of experience is appreciating. McKinsey’s human-capital research estimates that skills acquired through work experience account for 40 to 43% of average lifetime earnings in the US, UK, and Germany [18] — an enormous, hard-to-replicate store of value. The resolution to the paradox: AI compresses the part of experience that is procedural know-how, and rewards the part that is judgment, pattern recognition, and knowing what to do when the situation is off-script. The senior analyst from the opening is not threatened because she has experience; she is threatened only to the extent her experience was stored as procedure rather than as judgment. The reallocation, for an experienced professional, is to deliberately convert the former into the latter.

A diagnostic: which of your assets is which

Before you can reallocate, you have to mark your holdings to market — and most professionals have never audited their own capital with any honesty. The useful exercise is to take the work you are known for and sort it into three buckets. The first is procedural: tasks that follow a repeatable recipe, where a competent output can be specified in advance. These are your depreciating assets, whatever they once cost you to acquire. The second is judgment: the calls you make when the situation is ambiguous, the inputs are incomplete, and there is no recipe — which risk to take, which exception to grant, which stakeholder’s objection actually matters. The third is relational: the trust, credibility, and standing that let you mobilize other people. The second and third buckets are where your value is migrating.

A simple test sharpens the sort: for any given thing you do, ask whether a capable colleague with a good AI tool could produce a near-equivalent result tomorrow. If the answer is yes, that is a commodity in waiting, no matter how central it has been to your identity. The discomfort most experienced professionals feel here is that their proudest competence — the model, the brief, the analysis they are famous for being fast at — often lands in the first bucket. That is not a reason to despair; it is the whole point of the audit. You cannot reallocate capital you refuse to admit is depreciating. The professionals who adapt are the ones willing to look at the asset they are proudest of and ask, soberly, what it is still worth.

Reallocating your portfolio

None of this argues for abandoning your field, and that is the distinction worth drawing. Reallocation is not reinvention. You are not liquidating your portfolio and buying a new one; you are rebalancing toward the assets whose price is rising. In practice that means three moves. First, stop over-investing in the codifiable. You still need procedural fluency deep enough to understand the underlying logic and catch the machine’s confident errors — that foundation is exactly what lets the senior analyst verify what the junior’s copilot produced — but effort spent getting marginally faster at a task AI now does in a fraction of the time is capital poured into a depreciating asset. Second, deliberately build the complements — the judgment to direct the tool, the rigor to verify it, and the human skills that become the binding constraint once execution is cheap. Third, convert tacit experience into legible judgment: the pattern recognition that lives in your head is your most valuable holding, but only if you can apply it visibly and others can see you do it.

The organizational backdrop makes the case urgent. Employers themselves name skills gaps as the single biggest barrier to transformation, cited by 63% of them [19] — which means the capabilities you reallocate toward are exactly the ones the market is most starved for. The window to rebalance from a position of strength, while your current expertise still commands a premium, is the same window in which everyone else is also deciding whether to move. The hardest part is often not building the new capital but making it visible — naming the judgment and human capacity you have accumulated so the people who decide your trajectory can actually see it. (This is the work tools like Persona Map are built to support.)

The repricing is not a forecast; it is already on the tape. The professionals who thrive through it will not be the ones with the most skills, nor the ones who learned to prompt a chatbot fastest. They will be the ones who looked at their own portfolio honestly, recognized which assets the machine was making cheap, and moved their capital — their time, their attention, their deliberate practice — into the things that compound precisely because a machine cannot. Your career capital did not disappear when the regime changed. It got repriced. The only question is whether you reprice with it.

Sources

1. World Economic Forum. “Future of Jobs Report 2025: 78 Million New Job Opportunities by 2030 but Urgent Upskilling Needed.” 2025. Link
2. LinkedIn Economic Graph. “Work Change Report.” 2025. Link
3. World Economic Forum. “Future of Jobs Report 2025” (press release). 2025. Link
4. McKinsey & Company. “The Economic Potential of Generative AI: The Next Productivity Frontier.” 2023. Link
5. McKinsey & Company. “The Economic Potential of Generative AI: The Next Productivity Frontier.” 2023. Link
6. World Economic Forum. “Future of Jobs Report 2025 — Skills Outlook.” 2025. Link
7. PwC. “AI Linked to a Fourfold Increase in Productivity Growth” (2025 Global AI Jobs Barometer). 2025. Link
8. PwC. “AI Linked to a Fourfold Increase in Productivity Growth” (2025 Global AI Jobs Barometer). 2025. Link
9. PwC. “AI Linked to a Fourfold Increase in Productivity Growth” (2025 Global AI Jobs Barometer). 2025. Link
10. Lightcast, via PR Newswire. “AI Skills Command 28% Salary Premium as Demand Shifts Beyond Tech Industry.” 2025. Link
11. Anthropic. “The Anthropic Economic Index.” 2025. Link
12. Dell’Acqua et al. (Harvard Business School / BCG). “Navigating the Jagged Technological Frontier” — “How People Create and Destroy Value With Gen AI.” 2023. Link
13. Live Science (reporting Microsoft / Carnegie Mellon research). “Using AI Reduces Your Critical Thinking Skills, Microsoft Study Warns.” 2025. Link
14. McKinsey Global Institute. “Skill Shift: Automation and the Future of the Workforce.” 2018. Link
15. America Succeeds. “Durable by Design.” 2025. Link
16. World Economic Forum. “Future of Jobs Report 2025 — Skills Outlook.” 2025. Link
17. MIT Sloan. “Workers With Less Experience Gain the Most From Generative AI.” 2023. Link
18. McKinsey & Company. “Human Capital at Work: The Value of Experience.” 2022. Link
19. World Economic Forum. “Future of Jobs Report 2025” (press release). 2025. Link