Why AI Workforce Transformation Starts With End Users

John Healy shares what happened when he handed AI to end users and told them they could not change a single system or policy.

AI Workforce Transformation Starts With the People Doing the Work

Most AI workforce transformation efforts die in the same place. Leadership recognizes something is broken, hands it to HR, then hands it to procurement, then hires a consultancy to produce a roadmap. Twelve months later there is a deck, a steering committee, and a talent process that works exactly the way it did before.

John Healy spent 30 years at Kelly Services, where he ran the Office of the Future of Work. He has watched that pattern from the inside more times than he can count. So when the CEO of a high-tech company with 30,000 workers globally gave him free rein to reimagine how the organization builds capacity, he inverted the standard approach entirely.

The CEO's framing was unusually honest. He had tasked HR with fixing it. He had tasked purchasing with fixing it. Nobody was fixing it, and his entire user community was frustrated. What he wanted was an outside view.

John's first decision was who would be in the room. Not HR. Not procurement. The end users, the people responsible for getting work done and accountable for the output.

"We specifically said we don't need HR or procurement in the room. We have all your policies. I want to hear from the end users." - John Healy

Two Constraints That Made a Six-Month Transformation Possible

The second decision was the one that actually unlocked the room. John set two ground rules before anyone spoke.

You cannot change any technology. No arguing for a new VMS, no getting rid of Workday, no replacing the procurement platform. And you cannot change any policy. The rules are what they are, and they exist for a reason.

On paper those constraints sound like they would kill the exercise. In practice they were the reason it worked. When nothing anyone owns is on the table, nobody has to defend a decision they made. The person who signed the VMS contract has nothing to protect. The person who wrote the contractor onboarding policy has nothing to justify. What remains is an honest inventory of where the current process actually fails.

Before the session, John loaded the company's own material into their private Copilot instance: every workflow the software runs on, every policy manual covering full-time hires, contractors, and outsourced projects, plus a test environment and their Lightcast labor market data. He brought two reference frames with him. Mercer's experience architecture concept, where buying a book on Amazon takes two clicks while 70 systems work invisibly underneath. And Josh Bersin's HR 2030 model, which puts the average organization at roughly 140 disconnected systems touching how talent gets engaged.

Then he put process flows on a screen, asked what works and what fails, and typed every answer straight into the model as people said it out loud. Periodically he would ask the model to turn the accumulated input into a product requirement definition, push that into Lovable, and put a working interface back in front of the room.

"Three hours later, we had iterated enough that we now had this series of interfaces that would allow people to just kick off a project and get it done." - John Healy

Known Talent, Unknown Talent, and Why Trust Is the Real Constraint

The most valuable output of the workshop was not the software. It was a distinction the users arrived at themselves, one that no workforce taxonomy had handed them.

There is known talent and there is unknown talent.

Known talent might be a full-time employee whose skills and interests already sit in Workday. It might be a contractor who has delivered before, or a free agent already in the system with a profile. When a manager starts a project, that is where they want to look first, because the trust already exists.

Unknown talent is where everything slows down. A new contractor triggers an entirely different onboarding path. A new full-time hire means budget approval, headcount, and a posting. That friction is the organization buying insurance against someone it cannot vouch for.

"Known is an element of trust. If I have that, I can act in two clicks. If I don't have that, I can't." - John Healy

This reframes contingent workforce management in a way most operating models miss. The strategic question is not how fast your requisition moves. It is how large your pool of known talent is, and whether you are doing anything deliberate to grow it. That means networking in communities before you have a need, and it means treating external partners as a source of trusted people rather than as a line item to squeeze. As John puts it, procurement should absolutely control price, but not at the cost of eroding trust the organization is trying to build.

Why Rogue Spend Is a User Experience Problem

Every enterprise has contractor spend running outside the sanctioned process, and almost every enterprise treats it as a compliance failure. John's diagnosis is different, and the fix follows directly from it.

"All this rogue spend, why? Because this system doesn't work for me. I'm going around it. Well, give me a system that works for me, and I won't go around it." - John Healy

Nobody sets out to violate procurement policy. They set out to get something done. When the approved path takes six weeks and the unapproved one takes an afternoon, people take the afternoon and quietly accept the risk. Responding with more controls makes the approved path slower, which makes the workaround more attractive. That is the loop most organizations are stuck inside.

What John built instead was a faster path that was already approved and entirely optional. Nothing was turned off. The old system still runs for anyone who prefers it. Adoption spread by managers telling other managers how they now get a project started. Meanwhile every friction point the workshop surfaced in the policies went back to HR, procurement, legal, contracting, and IT, better documented than it would have been had those functions sat in the room.

Where Human Cloud Fits

What that company built by hand is a two-click front door to trusted talent. Be precise about what it required: a CEO willing to fund an outside view, an IT team willing to open a test environment, six months of iteration across workshops in Toronto, Singapore, and Durham, and an engineering effort to make vibe-coded prototypes hold up at enterprise scale.

Most organizations do not have that. They have the same frustrated end users, the same 140 disconnected systems, and no appetite for a six-month internal build.

That gap is the reason Human Cloud exists. We aggregate and verify solutions across 1,000+ workforce platforms, automate the discovery and compliance work that makes unknown talent slow, and give business teams a path from need to engaged partner in minutes rather than months. The known versus unknown problem John's users identified is exactly what verified performance data, business cases, and kudos are built to collapse: they turn a vendor you have never used into one you have grounds to trust.

John did not describe Human Cloud in that workshop. His users described the same destination independently, which is the strongest validation the thesis has received. The difference is that they had to build it, and you do not.

The Bottom Line

AI workforce transformation does not fail because the technology is immature. It fails because it gets handed to the functions that own the existing process, who are structurally unable to indict it. John Healy's result came from routing around that, and the constraints he imposed, change no technology and change no policy, are what made the room safe enough to be honest.

None of this removes the need for trust. It relocates it. Trust is still what separates a two-click hire from a six-week requisition, and it is still built through relationships, track records, and communities. What AI changes is that the proof can finally be made visible, and the workarounds built to compensate for its absence can finally be retired.

---

About John Healy

John Healy is Chief Executive at WHRRR.WORK. He spent 30 years at Kelly Services, where he ran the Office of the Future of Work, and now advises companies and governments on radically rethinking how people connect with work.

Listen to the full episode: Human Cloud Podcast on Spotify

---

This article was adapted from the Human Cloud Podcast. Subscribe wherever you get your podcasts.