This is the final post in a series exploring the idea of skills you earn vs. skills you learn. The previous posts can be found in my archives.
If you’ve followed the series from the beginning, you’ll have noticed that I’ve been circling the same problem from different angles for the best part of four posts. The enterprise graph could see what people did, but couldn’t carry it anywhere. The TED talk argued that analytics should serve the individual first, but the infrastructure to make that real didn’t exist. Post 4 traced why closing that gap took a decade — incumbency, business model inertia, the chicken-and-egg problem of network adoption, and eventually the forcing functions that broke the deadlock.
This post is where all of that arrives somewhere.
The problem, plainly stated
We have two categories of professional skill, and we treat them as if they were one.
Skills you learn are acquired through study, training, and certification. They are point-in-time — a course completed, an exam passed, a credential issued. They are well served by existing infrastructure: transcripts, certificates, Open Badges, digital diplomas. The system for documenting what someone learned is imperfect but functional.
Skills you earn are different in kind, not just degree. They accumulate through doing — through the project that shipped under pressure, the client relationship that held through a difficult delivery, the team that succeeded because of how someone led it. They are not point-in-time; they compound. And they are almost entirely invisible to any formal system.
The project manager who has successfully delivered six enterprise cloud migrations has something no course completion can represent. Not a qualification — a track record. A body of evidence, accumulated across a career, that tells you something real about who this person has become and what they can actually do. The problem is that this evidence exists nowhere that a new employer, a client evaluating a proposal, or an AI matching system can see and trust.
That’s not a small gap. It’s the gap between what a person is worth and what the market can verify they’re worth. For some people — the quiet contributor, the excellent deliverer who doesn’t self-promote, the person whose best work was always for someone who can’t name names — it’s the gap between the career they deserve and the career they get.
What changes when you hold your own credentials
The worker who carries a digital wallet of verified, issuer-signed proof-of-work credentials is in a fundamentally different position from the one carrying a CV and hoping a reference call comes through.
Each credential in the wallet is a micro-reference — a small, specific, cryptographically signed attestation from someone who witnessed the work directly. Not one or two references assembled at the end of a tenure. A continuous accumulation, across every project, every client, every outcome that someone with authority chose to attest to. The traditional reference is a blunt instrument: high friction, end-loaded, easy to game, dependent on the goodwill of a single person. Micro-references are granular, continuous, issuer-verified, and portable. Their accumulation over a career is a richer signal than any single reference call.
There’s a timing dimension here that matters. For most of the past decade, the argument for verified proof-of-work credentials ran ahead of the infrastructure to consume them. A recruiter reading a resume has a practical limit — a few pages, a phone call or two. But AI-powered hiring and matching systems have no such constraint. They can analyze vast amounts of structured, trusted data about an individual simultaneously — skills, outcomes, tenure, client feedback, delivery context — and make genuinely data-driven decisions at a scale no human reviewer could match. The catch is that this only works if the underlying data is trustworthy. Self-reported CVs fed into an AI matching engine produce faster versions of the same biased, gameable output. Issuer-signed, cryptographically verified credentials fed into the same engine produce something qualitatively different: decisions grounded in evidence rather than assertion. AI hasn’t just created demand for better credential data — it has created verifiers capable of consuming it at scale. That changes the network dynamics considerably. The chicken-and-egg problem looks different when one side of it is an AI system that can process thousands of verified credentials the moment they exist.
Portability is not incidental. It’s the point. The credential travels across employer boundaries when you change jobs. It travels across sector boundaries when you move industries. It travels across national boundaries when you work internationally. You did the work in one context; the evidence of it is yours to carry into every context that follows. That’s a new kind of career capital — and it changes the power dynamic between workers and the institutions that employ them.
Why employers should want portability too
The predictable concern from employers at this point is about portability working against them. If we invest in building rich proof-of-work credentials for our people, don’t we just make them more attractive to our competitors?
The answer is yes — and that’s not the right frame.
When people decide which car to buy, resale value is a real factor in the calculation. Not because everyone plans to sell immediately, but because high resale value signals quality and broad market confidence. An employer who helps you build a portable, verifiable career record is offering the professional equivalent. The best people — the ones every organization wants — will increasingly factor this into decisions about where to work. Not just salary, not just learning opportunities, but the career capital they’ll carry with them when they move on. Companies that issue rich, governed credentials are offering something that a competitor, offering a payslip and a LinkedIn endorsement, cannot match.
The network effects run deeper than individual recruitment. An organization whose people carry strong, verifiable credentials becomes more trusted by its clients. Its proposals are more credible when the track records behind them are independently verifiable rather than self-asserted. Its AI matching systems become genuinely intelligent when they’re operating on issuer-signed proof-of-work rather than self-reported skills matrices.
The portability risk is real but second-order. The trust and talent benefits are first-order.
The confidentiality inversion
One objection to proof-of-work credentials in professional services has always been confidentiality. The client is confidential. The deal terms are confidential. You can’t issue a credential that names them.
This turns out to be a red herring — and understanding why is important, because the same logic applies across almost every sector where proof-of-work credentials matter.
It’s worth being explicit about this, because it’s not obvious until you understand how trust works in a verifiable credential system. The trust doesn’t live in the content of the credential — it lives in the governance around who issued it. A consultant can carry a credential that says “led a major digital transformation program for a global telecommunications company, delivered on time and budget, rated outstanding by the client account team” without naming the client, the country, or the contract value. No confidential information is disclosed. But the credential is cryptographically signed by a verified issuer — the employer, operating within a governed trust framework — and that signature is what makes it trusted by any relying party in the network, including the AI matching agent deciding who should be staffed on the next engagement.
Trust comes from governance, not content. The confidentiality constraint that has always been used to explain why proof-of-work can’t be issued turns out to be irrelevant to the question of whether it can be trusted.
California and the public sector proof point
The workforce credentialing problem I’ve been describing in enterprise terms has an exact structural parallel in public sector skills programs — and it’s being addressed directly.
California’s Career Passport procurement is building the infrastructure to give community college learners a holder-controlled, verifiable record of what they’ve learned and what they’ve done — across academic credentials, vocational training, apprenticeship milestones, and work experience. The population it serves is one where the gap between real skills and visible skills is widest: workers whose contribution is genuine but whose credentials are fragmented, institutional, and invisible to the labor market systems that could connect them to opportunity.
The worker whose skills are real but invisible is the infrastructure failure the Career Passport was built to fix. That sentence applies equally to the community college learner in Fresno and the consultant whose best project was for a client they can’t name publicly. The problem is the same. And so is the architecture of the solution — including the answer to the confidentiality question.
What we’re asking for
The infrastructure exists. The standards are in place. The regulatory framework — at least in Europe, and increasingly elsewhere — is moving. What’s needed now is adoption: the employers, training providers, and public programs willing to issue credentials that mean something, into wallets that workers actually hold and use.
That requires resolving the business model questions that Post 4 named as still open. It requires incumbent players deciding that participation in the network is more valuable than defense of the existing model. It requires the chicken-and-egg problem to be broken by enough anchoring use cases — the California Career Passports, the enterprise proof-of-work pilots, the government identity wallets — that the network reaches the density where it becomes self-sustaining.
None of that is certain. Transitions of this kind never arrive on schedule, and they rarely arrive cleanly. But the direction is no longer in doubt. The question is how long it takes, and who moves early enough to shape it.
If you’ve been reading since Post 1, you’ll know I’ve been thinking about this since 2013. I’m under no illusions about the distance still to travel. But I remain convinced — for the same reasons I was convinced then, and better-evidenced ones now — that the model that serves the individual first will, in the end, serve everyone better.
You did the work. Now prove it.
Marie Wallace leads the Digital Identity Innovation practice at Accenture. She has been writing about the human side of data at allthingsanalytics.com since 2011. All opinions expressed are her own.
Marie's Ramblings & Ruminations 