I stopped being John Henry

For most of my career, my desk told the same story every Friday afternoon. A stack of folders. Each one with a Post-it note on top, written in my handwriting: "Waiting on the data section." "Sent to legal Tuesday, still nothing." "Need program review before Monday."

I spent thirty years in Missouri state government. Child support program administration, policy work, fatherhood initiatives. I had a team down the hall, a legal office on the org chart, a data analyst three doors down, and the full weight of a state agency's infrastructure behind every memo. 

And still, the rhythm was the same. Two days here. Three days there. A reviewer juggling four other priorities. By the time a draft came back, I had already moved on to two new projects and had to spin myself back up to remember why I sent it out in the first place.

The bottleneck was never me. It was never even them, exactly. It was the wait. The handoff. The calendar drift between the moment a thing was ready in my head and the moment it actually got out the door. I billed 30 years inside that rhythm and assumed it was simply how the work moved.

Then I left state service, started Ginwright Professional Services, and saw something I should have seen earlier. The bottleneck was structural. State government had it. The Am Law firms have it. Every consulting shop and grant-writing operation has it. The Post-it notes on my desk in 1996 were the same Post-it notes that show up on every knowledge worker's desk in 2026. We just call them Slack reminders now.

That is when I stopped being John Henry.

The story of John Henry is one of the most American things we have. A Black steel-driver takes on a steam-powered drill through the rock of the Big Bend Tunnel. Two ten-pound hammers, one in each hand. He wins. And then he dies, hammer in hand, heart given out from the effort. We tell that story like a tragedy of dignity. But sit with it. What did John Henry actually win? Nine extra feet of tunnel. The railroad still got built. The steam drill still came. The men with hammers still lost their jobs.

Knowledge workers have been swinging hammers at the same kind of contest for the last few years. Every program manager who watched an artificial intelligence agent draft a proposal in 20 minutes that used to take three days, and then quietly told themselves the AI version "wasn't quite right" so they could redo it by hand, has been auditioning for the same legend.

Here is the part most AI articles get wrong. You do not type a prompt, hit enter and download a finished proposal. That is not how this works. I can spot a prompted-and-downloaded proposal the moment I open it. So can every reviewer who has read more than ten of them. The structure is generic. The voice is wrong. The specifics are missing or hallucinated.

What I do is different. I sit with the AI agent the way I worked with the best staff I ever managed inside the state. I give it the foundation: the client, the funder, the regulatory context, the program model, the budget logic, the political read of the room. I read the first draft. I push back. I rewrite sections. I bring in the parts that only thirty years of working inside the system would tell you matter. 

The agent does not know which sentences will catch a state reviewer's eye. I do. The agent does not know which fatherhood program model has moved the numbers in a IV-D environment. I do. The agent amplifies what you bring to it. Bring 30 years, you get 30 years. Bring nothing, you get nothing dressed up to look like something. That distinction is the entire ballgame.

Eight to ten focused hours on a single proposal now replaces three weeks of calendar drift. The working time was always ten hours. The other two weeks and three days were waiting. The work didn't get cheaper. The wait disappeared.

Three things shifted when the wait disappeared, and they would have shifted whether I was a startup or still inside the state. The calendar collapsed. The St. Louis Federal Reserve estimates the average AI user saves about 5.4% of their work hours. For document-heavy work the published numbers run far higher: an Am Law 100 firm reported a 50% to 67% reduction in document review time, hitting accuracy at or above human first-level reviewers. 

The margin showed up. Every two-day wait used to eat into a fixed-fee deliverable or stretch an hourly engagement past what the client wanted to pay. State agencies sit on the same leak with larger budgets. The judgment finally caught up to the output. For my entire state career, my judgment was throttled by the production layer underneath it. The moment that layer ran at the speed of my decisions, the work got better, not just faster.

There is a quiet pride in knowledge work that says the value is in the doing. The drafting, the formatting, the citation pulling, the red-line cycles. That pride is the John Henry fallacy. It mistakes the swinging of the hammer for the building of the tunnel. The C&O Railroad needed a tunnel. It did not need a steel-driver. The client needs a winning proposal. They do not need a three-layer human review cycle. McKinsey sized the long-term corporate AI productivity opportunity at $4.4 trillion. That number is large because the fallacy is so widely held.

I am not the steel-driver in this story anymore. I am also not the engineer who walks away while the drill runs unattended. I am the contractor who hires the machine, knows how it works, sits next to it, corrects it, and finishes the tunnel ahead of schedule. The judgment is still mine. Thirty years are still mine. They just show up faster now.

The hammer is still on my desk. I keep it there to remind me what the work used to cost.

John Ginwright is the CEO of Ginwright Professional Services, a Missouri-based consulting firm specializing in child support policy, fatherhood program development, government grants and RFP response and AI-augmented program design. He spent 30 years in Missouri state government before founding the firm. He is currently enrolled in MIT Professional Education's Generative AI program and is a Missouri-certified MBE.