The hundred-year machine
For one hundred years the path to partner was simple. You did the grunt work until you earned the judgment.
You know this machine intimately. A junior does the manual data entry and the messy first draft. A senior bleeds red ink over the page. The junior fixes it. After a few thousand cycles, the junior stops needing the check. They develop a quiet instinct for the work.
The grunt work was never just labor. It was the tuition you paid to learn the shape of a good deal.
The system worked perfectly. It produced capable leaders and built reliable firms. Every partner walking the halls today learned their trade through this exact mechanism.
The agent takes the reps
The data is completely clear. AI agents do those reps now.
The month-end close just went from five days of human grinding to eight hours of automated processing. A regional CPA firm recently cut their tax return prep from four hours of manual entry to thirty minutes of human review. The entire structure of professional services is changing overnight.
The old model, where the junior does the work and the senior reviews, evolved to a new model where the agent does the work, and the human supervises.
The traditional entry-level job is vanishing. The organizational chart is flattening into a diamond. It is getting very thin at the bottom and widening out in the middle.
The trap most firms will fall into
This new model is undeniably faster and vastly cheaper. I say this as a builder completely invested in this shift. At Vixure, we build these exact automated systems for B2B SMBs every single day. The savings are massive and immediate. You can double your output without adding a single desk. That is exactly why the trap works so well.
The short-term incentive is overwhelming. It makes complete financial sense to hand the grunt work to the machine. But this is the question that keeps me up at night.
If a twenty-five-year-old never does the grunt work, where does their judgment come from in ten years?
You cannot supervise something you never learned to do yourself. Supervision without earned judgment is just watching. A manager without battle scars will accept a machine's confident hallucination. They will rubber-stamp it. The client will take the hit.
The deliberate alternative
Most owners cannot answer two critical questions. Where is AI actually saving hours in your firm? And where are those saved hours going? That is the whole assessment. Almost nobody has run it.
You must keep the AI and deliberately change the reps. The reps were never sacred. The judgment was. You need to design entirely new reps that build judgment faster than the old grunt work did.
Here are three moves you can run this quarter:
- Argue with the agent. Make your juniors defend the AI's output against your critique. Make them find the hidden holes. Review their reasoning instead of just reviewing the final work product.
- Run the delta. Have a junior do one matter entirely by hand each month. Compare it against the agent's version. The delta between the two is the lesson.
- Buy client exposure. The hours AI gives back should buy client exposure years ahead of the old schedule. Put juniors in the room earlier. They can build judgment from watching real decisions and learning how to navigate difficult conversations.
Who owns the next decade
Before you try any of these moves, take stock. Map where AI is doing work in your firm today. Map who checks it. Map what your people stopped doing because of it. We run this as a free AI Readiness Assessment with firms every week. You can also run it yourself in an afternoon with a whiteboard.
The old ladder is not coming back. The firms that learn to grow judgment in an AI world will own the next decade. The ones that just cut the bottom of the ladder will run out of people who can climb it. We have to build a new path forward.
How is your firm making sure the next generation still earns its judgment?
P.S. If you want a clear picture of where AI is really working in your firm and where the hours are going, the AI Readiness Assessment is free and takes one conversation. No pitch. Just the map.
Sources and data
- MIT Technology Review, "Rethinking organizational design in the age of agentic AI" (May 26, 2026).
- Intuz / DualEntry / Karbon, 2026.
- Insightful Accountant case study, 2026.
- Box / McKinsey / LinkedIn, "20 agentic AI jobs" (2026).
- Anthropic Claude for Legal launch / Fortune (May 2026).