The high-agency mindset: what BCG's viral phrase gets right, and what – Good Machine

TL;DR

BCG behavioural scientist Julia Dhar told Business Insider on 26 June that workers thriving in the AI era share what she calls a "high-agency mindset" — belief that their actions matter, coupled with pragmatic optimism. The phrase is trending across HR and leadership feeds this week.
Underneath the phrase sits BCG's 2026 AI at Work Research: 72% of workers say the skill expectations of their role have already changed; 88% believe they'll need major upskilling within five years.
The framing is landing simultaneously with hard structural signals: Deloitte's 2026 Global Gen Z & Millennial Survey (Singapore cut, released this week) showing only ~25% of Gen Zs and ~20% of millennials in Singapore now chase fast-paced career growth; India ranked 74th in workforce readiness by the QS World Future Skills Index 2027; Singapore's new Skills and Workforce Development Agency (SWDA) officially standing up on 1 July; and Africa's ALX (founder Fred Swaniker) profiled by CNN on 1 July as it scales pan-African tech reskilling.
The trap: "high agency" flatters the reader into thinking mindset is the primary variable. It isn't. The primary variable is whether the ground under your job is moving — and it is, unevenly, across continents.
What actually works: five small, verifiable behaviours that build agency in practice, not in vibe. They're at the bottom of this piece.

The framing that went viral

Here is Dhar's line, quoted across a half-dozen outlets in the last seven days:

"The more you take action, the more that it ends up expanding the surface area of opportunity."

It is the kind of sentence that survives being screenshotted. It is also the kind of sentence that means very different things to a Bengaluru graduate whose first-year analyst role is being consolidated into an agent workflow, a São Paulo call-centre worker whose employer has just switched to a voice model, and a Sydney manager deciding which two of her twelve reports to keep. All three of them, technically, need "high agency." Only one of them has the runway to practise it without material consequence.

That gap — between the phrase and the ground — is where this story lives.

What actually happened this week

Three things landed at once, and they compound.

One. Business Insider published Dhar's "high-agency mindset" interview on 26 June, followed on 2 July by a companion piece surveying talent leaders at BCG, EY, and Cognizant. The through-line: soft skills — adaptability, curiosity, judgement — now outrank technical credentials for entry-level hiring. BCG's Brian Myerholtz, global head of talent acquisition, told BI: "AI has raised the bar for what [judgement and problem solving] entails."

Two. Forbes ran a Dhar follow-up on 2 July arguing that most organisational change fails behaviourally, not strategically — because executives "default to optimism employees don't share." The subtext: agency at the individual level is being asked to compensate for agency deficits at the organisational level.

Three. Adjacent evidence that the ground is moving, from three continents in a single week:

Asia — Deloitte's Global 2026 Gen Z & Millennial Survey (Singapore cut) found only about a quarter of Gen Zs and a fifth of millennials want fast-paced career growth; roughly half have postponed major life decisions on financial grounds. India's QS World Future Skills Index 2027, released 28 June, ranks India 1st in economic capacity for an AI-driven future — and 74th in workforce readiness. The gap is the story.
Africa — CNN's 1 July feature on Fred Swaniker's African Leadership Xcelerator (ALX) charts explicit continent-scale tech reskilling — software engineering, data, and now film, music, gaming, animation — pitched at remote-work-ready talent.
Europe/Global — The Guardian's 26 June piece on European heatwaves estimates a four-day heat event could shave 1.5 percentage points off UK quarterly labour productivity and up to 2 in the rest of western Europe. Agency, meet climate.

Put those alongside a Forbes piece from 27 June — AI Is Eroding Your Organization's Intellectual Capital — which cites a Microsoft/Carnegie Mellon study of 319 knowledge workers finding that higher confidence in AI correlates with less critical thinking applied to checking it, and the shape of the moment becomes clear. Everyone agrees the workforce needs to change. Nobody quite agrees on who is responsible for the changing.

The framework, honestly examined

Dhar's "high-agency mindset" is not new — it borrows from decades of work on self-efficacy (Bandura), locus of control (Rotter), and, more recently, Angela Duckworth on grit and Carol Dweck on growth mindset. What is new is the packaging, and the timing: it lands exactly when workers are being told, simultaneously, that (a) their skills are depreciating faster than at any point in the last 30 years, and (b) the answer is inside them.

The framework is genuinely useful when you strip the packaging off. Broken down, it's three behaviours:

Belief that your action changes the outcome. In psychology this is internal locus of control. Empirically robust; correlates with career resilience across decades of longitudinal data.
Pragmatic optimism. Not certainty that things will be good — expectation that acting will produce information.
Bias toward small, cheap, reversible moves. Dhar's "surface area of opportunity" line points at this without quite naming it.

Those three are real. What the viral framing tends to skip is the fourth thing, which is that all three of them are dramatically easier when you have a financial cushion, a professional network, and an employer who invests in your development. High agency is not evenly distributed because the raw materials for it are not evenly distributed.

Which is why the phrase risks becoming what psychologists call a dispositional attribution error dressed as career advice. If your career is thriving, you have a high-agency mindset. If it isn't, you don't. The reasoning is circular and the diagnosis is unfalsifiable.

Who benefits from this framing being everywhere

Follow the incentives.

Consultancies benefit — a mindset problem is a training problem, and training is what they sell. BCG has a Behavioral Science Lab; the framing directly serves its product.
Employers benefit — if the story is that thriving workers have the right mindset, the story is not that employers have under-invested in reskilling. Deloitte's own data shows the gap: workers say they need major upskilling; most say their employers aren't providing it at the required pace.
Media benefit — a punchy, individual, forwardable phrase outperforms a systemic story every time.
High-performing individuals benefit — the framing tells them they earned it.

Who doesn't benefit: the worker whose employer has quietly stopped funding L&D, whose team has been slimmed, and who is being told the differentiator is inside them.

None of which makes the underlying advice wrong. It makes the framing incomplete.

Cross-layer scan — the connections most pieces are missing

Education systems — India's #1-in-economic-capacity / #74-in-workforce-readiness split is the sharpest globally, but not the only one. Singapore's SWDA merger of SkillsFuture and Workforce Singapore under one ministry is a bet that agency has to be built at the system layer, not the individual layer. Watch whether other governments copy the structure.
Talent mobility — India's 28 Mobility Partnership Agreements (Jaishankar, 30 June) are turning skilled labour into a foreign-policy instrument. The individual "high-agency" narrative and the state-level "mobility pact" narrative are the same underlying trend viewed from opposite ends of the telescope.
AI-and-critical-thinking coupling — the Microsoft/Carnegie Mellon finding that AI reliance erodes critical thinking is directly downstream of "agency." If action expands opportunity, unexamined action expands opportunity for the tool, not the worker.
Climate — the European heatwave productivity data is a reminder that agency isn't only social. A workforce that can't cool its offices can't out-mindset a heat dome.

The quieter story

The story most outlets are not writing this week is that the entry-level job is disappearing faster than the systems designed to replace it. The AI Journal's 1 July piece on EasyA's Dom and Phil Kwok (Planet Classroom) puts the frame bluntly: "The entry-level job as we know it, I think, is over." If they're even directionally right, then a decade's worth of "start at the bottom, learn by doing" career logic breaks. Which means "high agency" isn't a mindset problem for early-career workers. It's an infrastructure problem — how do you build a first ten thousand hours when the first thousand have been automated?

That's the story to watch through the second half of 2026.

What this actually means for you

Recommendations are addressed to the general public — anyone whose work touches a screen, a customer, or a decision. Where the honest answer is "the useful move depends on your position," it's tiered.

If you're early-career (0–5 years experience, any geography)

Do the thing that produces public evidence. A shipped side project, a written analysis, a talk given, a repo maintained. Recruiters — including at BCG, EY, and Cognizant per this week's reporting — now index heavily on demonstrable curiosity and judgement. A public artefact beats a résumé line.
Deliberately un-automate one part of your work per quarter. Do it by hand. This is the direct counter-move to the Microsoft/Carnegie Mellon finding. You are protecting the critical thinking muscle that will matter more, not less, as AI tools proliferate.
Find a specific mentor and ask a specific question every fortnight. "How would you have thought about this?" is disproportionately effective and disproportionately underused.

If you're mid-career (5–15 years) in a role adjacent to knowledge work

Audit which parts of your job the current generation of AI tools already does better than you. Not to panic — to reallocate. If your value was in first drafts, your value now is in judgement about the draft. Learn to articulate that shift out loud, in performance reviews and 1:1s.
Pick one adjacent domain to become dangerously literate in over 90 days. Not to change jobs — to expand what BCG calls "surface area of opportunity." Empirically, career resilience correlates with span more than with depth alone.
Track your own upskilling as a line item. Time booked, hours completed, artefacts produced. If your employer isn't funding it (per Deloitte, most aren't at the required pace), you are running a personal training programme; run it like one.

If you manage people

Stop asking your team to demonstrate "high agency" while withholding the conditions that make it possible. Concretely: protected learning time, budget for external training, clear guidance on which AI tools are sanctioned, and psychological safety to admit skill gaps.
The behavioural on-ramps matter more than the vision. Dhar's own Forbes piece names this: change fails because leaders assume communication equals adoption. Build the on-ramp.

If you're a policy-maker or L&D leader

Watch Singapore's SWDA closely. Whatever its outcome, it is the most consequential experiment this year in whether workforce agency can be built at the system layer rather than the individual layer. Success or failure both instruct.
Fund the boring middle. The viral energy is on entry-level disruption and executive AI strategy. The workforce that will need the most support — mid-career, mid-skill, mid-life — is where the least attention is going.

If none of the above applies to you

Then the honest answer is: this story is background noise for you today. It will be foreground within 24 months. The single useful move is to notice which parts of your current work depend on tasks a competent AI system already does at 80% quality, and start compounding the skills that don't.

The uncertainty ledger

How viral, really? "High-agency mindset" is spreading in HR and leadership circles this week; whether it graduates to mainstream mental furniture is unclear. The Deloitte and QS data around it is the more durable substrate.
Whether the BCG framing survives its own success. Frameworks that become buzzwords tend to lose their operational meaning within 18 months. If "high agency" ends up on a motivational poster, its useful life will already be over.
How fast the entry-level job actually erodes. The Kwok-brothers framing is confident; the empirical evidence is directional but incomplete. If they're right, the recommendations above tighten sharply. If they're 40% right, they still tighten.
Whether Singapore's SWDA delivers. Too new to judge. Twelve-month watch item.

Bottom line

The "high-agency mindset" is a real, useful, portable idea. It is not a substitute for the fact that the labour market is being restructured by capital, technology, and demography at a pace the average worker did not consent to and cannot outrun by attitude alone. Practice the behaviours. Refuse the individual-blame framing. Ask, out loud and often, what your employer, your government, and your educational institutions are doing to meet you halfway. Agency you have to build entirely on your own is not a mindset — it is a burden being redescribed as a virtue.

Sources

Tier 1 — Business Insider, "Workers with this mindset are thriving in the AI era, says BCG leader," 26 June 2026; Business Insider, "Leaders from BCG, EY and more share the qualities they look for in entry-level talent," 2 July 2026; Reuters, "Global financial firms pivot to South Korea…," 30 June 2026; The Guardian, "Too hot for work: why extreme heat is a threat to Europe's productivity," 26 June 2026; Firstpost (S. Jaishankar, Human Resource Mobility Forum), 30 June 2026; CNN, "'Africa is the next wave of global talent': ALX / Fred Swaniker," 1 July 2026.
Tier 2 — Forbes, "Your Employees Are The Customers Of Change" (Julia Dhar interview), 2 July 2026; Forbes, "AI Is Eroding Your Organization's Intellectual Capital," 27 June 2026; The Times of India, "India Number 1 in economic capacity, 74th in workforce readiness: QS," 28 June 2026; HR Executive, "How Gen Zs and millennials are rewriting the rules of career success" (Deloitte 2026 Global Survey, Singapore cut), 1 July 2026; AsiaOne, "Skills and Workforce Development Agency (SWDA) launch," 1 July 2026.
Tier 3 — The AI Journal / PRNewswire, "AI Is Forcing a New Way Into the Workforce" (EasyA / Planet Classroom), 1 July 2026; Consultancy.uk, "The power of independence in an era shaped by complexity and AI," 26 June 2026.
Underlying research referenced — BCG 2026 AI at Work Research; Deloitte Global 2026 Gen Z & Millennial Survey; QS World Future Skills Index 2027; Microsoft & Carnegie Mellon study of 319 knowledge workers; Harvard Business School study on AI-augmented expert judgement (as cited by Forbes, 27 June 2026).