July 2026 · 7 min read
Hands Went Abundant. Judgment Stayed Scarce.
Resourcing prices hands, the one input that just went abundant. It ignores the two that decide whether the work is any good: judgment, which takes a decade to grow, and attention, which is being spent by default. Agents widened the gap, and no spreadsheet shows it.
Hand a junior real scope with no one beside them and you’ve built a fast car with an empty passenger seat. They’re told to bring alternative views, lean in, challenge the room, and then left to make decisions they have never watched anyone make. It’s the keys to a Ferrari when you’ve only ever driven a pickup, and no one riding along. When it goes wrong, the org reads the wreck as a verdict on the driver. It isn’t. It’s a verdict on the seat nobody filled, and the seat is empty for a reason that runs through every resourcing choice the company makes.
Resourcing prices the input that went abundant
Most resourcing conversations I’ve sat in count bodies. Headcount, allocations, fractions of an FTE spread across initiatives. The spreadsheet answers one question, who is assigned to what, and treats that as the whole of resourcing. It prices hands.
Hands are the abundant part, and more abundant now. What the spreadsheet never prices are the two inputs that decide whether the hands produce anything worth having: the judgment to make the hard decision when the work stops matching the plan, and the attention to hold a problem long enough to make it. Both are invisible in every allocation I’ve seen. Both are the actual constraint.
Two scarcities, and the plan sees neither
Judgment is scarce because it is slow to grow. It is built from years of watching decisions get made and learning which ones were wrong, and you cannot hire it on a quarter’s notice or prompt it out of a model. When you are short of judgment, you are short for a long time.
Attention is scarce in the opposite way: abundant in principle, starved in practice. Your senior people have judgment. What they lack is room, because their attention is being spent by default, scattered across everything that routes to them, with nobody deciding what it should skip. Attention can be redirected this quarter. Judgment cannot be built this quarter at any price.
The genuinely rare resource gets compounded by the merely unmanaged one. The org owns more judgment than it can spend, and spends less than it owns. A plan that counts hands sees neither problem.
Agents didn’t remove the work. They moved it.
Agents collapse the cost of routine execution. Google said in late 2024 that more than a quarter of its new code was generated by AI, then reviewed and accepted by engineers, and the number has climbed since. Read that sentence twice. The generation got cheap. The reviewing and accepting did not, and it moved onto the engineers.
That is the general shape. The work does not disappear. It moves. Somebody still has to decide whether the plausible output is the correct one, whether it should ship, whether the thing being built should exist. Microsoft’s 2025 study of knowledge workers found the same shift: as AI takes over execution, human effort moves toward verifying, integrating, and stewarding what the machine produced. Even craft moved up the stack: knowing a good draft from a plausible one is now most of what doing it well means. Agents converted hands-work into judgment-work, the one kind the org was already short of. The input that went abundant is the one the spreadsheet counts. The input that got scarcer is the one it ignores.
A junior with an agent is not a junior with a copilot. It’s the same fast car with a bigger engine, accelerating whatever direction the driver picks, silent on whether to hold back.
Picture what that looks like in practice. A junior points an agent at a credit eligibility rule, it drafts in an afternoon, the output looks right, and it ships. The rule starts issuing credits to accounts that shouldn’t qualify. Nothing alerts, because nothing is broken. Finance finds it a quarter later, and clawing money back from your own customers costs more trust than the credits cost cash. Every step was fast. The missing piece was five minutes with someone who had seen an action that can’t quietly be taken back and knew this was one. The empty seat costs more with the agent in the car, not less.
You may not have enough
Here is the part leaders skip, because it has no comfortable answer. You may not have enough judgment for the work you have committed to. Not spread thin. Not enough, because it takes years to make and you did not start early enough.
No reshuffle fixes that. Naming a mentor for every junior is impossible when there are not enough experienced people to name. Resourcing stops being the art of assigning bodies and becomes the art of rationing: deciding where scarce judgment gets spent, and admitting what you are choosing not to cover. The three moves that follow are what rationing looks like. All three sit inside a leader’s own span. None wait on the portfolio shrinking.
The seat does two jobs
A delivery seat puts senior judgment next to a decision the company cannot afford to get wrong. A development seat puts a junior next to a decision so they can watch it get made. The first protects this quarter. The second manufactures the judgment you’ll be short of in five years. Treat them as one kind of seat and you get the worst of both: seniors chaperoning work that didn’t need them, juniors alone on work that did.
Ration the delivery seats by the cost of being wrong. What does a confident, wrong answer cost here? The bet feeding the board number, the action you can’t take back, the decision that sets precedent: that’s where a senior gets pulled off everything else to sit. The internal experiment nobody steers by runs with the seat empty, and should.
Give juniors the keys on work the company can afford to crash. Driving is how you learn to drive, and recoverable work is where mistakes are tuition instead of losses. But driving alone never teaches the deciding. That’s what the development seat is for: scheduled time riding along on the hard decisions, as its own line in the plan. Growth is an allocation, not a residue. If the only time a junior sits with a senior is when delivery happens to require it, you are not developing anyone. You are occasionally being watched.
Count attention as load. A senior carrying four juniors and review duty on six agent-built systems is fully allocated, whatever the spreadsheet says, and the invisible version of that load is what turns your best people into the funnel. You don’t need portfolio authority to fix this for your own team. Put seat time and review time on the same page as project time, and when the page is full, say so in writing. The roles can be perfectly designed, and none of it holds if the people carrying the judgment have no accounted hours to spend it.
The cost that compounds
Judgment is not only spent at the seat. It is made there. The junior riding along on a hard decision, learning why it was hard, is how the next decade’s judgment gets built. Leave every seat empty and you aren’t just shipping worse work now. You are failing to manufacture the thing you are already short of, and you’ll be as short in three years, having run a generation of juniors past the point where anyone taught them to decide.
Attention is the lever you hold this quarter. Judgment is the one you grow across a decade. A resourcing plan that prices neither is not a plan. It is a count of the one thing that stopped being scarce, kept carefully up to date while both of the things that matter run out.