Automation can make a company stronger while reducing the number of people it plans to hire. “Jobs created” therefore measures one visible outcome, but misses whether a small company extended its runway, protected existing roles, or created capacity to grow.
At 2:14 PM on Tuesday, I reached that field in an awards application. Hours earlier, I had approved an automation that removed the need for our next operations hire.
The field expected a number. The honest answer required a paragraph.
When the old measure meets the new company
In 1975, Kodak engineer Steven Sasson built the first self-contained digital camera at the company’s research laboratory in Rochester, New York. The device was an early experiment, but the conflict it exposed was already clear.
Kodak’s business depended on film. Sasson had built a camera that did not use it.
The National Museum of American History documents the camera and Sasson’s work. His prototype did not prove that digital photography would replace film. It did, however, force Kodak to consider a possibility its existing commercial measures were poorly designed to value.
A digital camera could look unimpressive when judged by film sales. An automation can look unimpressive when judged by immediate hiring.
That was the problem sitting inside the awards form. It treated employment growth as a direct measure of company progress. Sometimes it is. A growing team can signal demand, investment and confidence. But for a founder working with limited runway, an avoided hire can also be evidence of a better operating decision.
The distinction matters. “We did not hire” can mean the company stalled. It can also mean recurring work was redesigned so the existing team could handle more volume without adding another monthly obligation.
One field cannot tell those stories apart.
The hire was a forecast, not a trophy
A planned hire often begins as a forecast about work.
Customer records will need updating. Documents will need checking. Someone will have to move information between systems, follow up on missing details, and tell the next person when a task is ready. Once those actions appear often enough, the company turns them into a role.
That translation can happen too quickly.
Before opening the role, I need to ask what kind of work created it. Does the task require judgment? Does it involve an exception whose cost justifies human review? Or are we paying someone to copy information, monitor a queue and send the same prompt repeatedly?
Automation changes the answer to the hiring question by separating those categories.
The point is not to remove people from work wherever software can perform a step. The point is to stop treating repeated coordination as proof that another full-time role must exist. A small company can preserve cash for engineering, customer support or distribution by removing work that never needed human judgment.
That choice still carries risk. A fragile workflow can fail quietly. An automated decision can cross an authority boundary. A system can process the ordinary case and mishandle the one customer situation that matters most.
I have seen the same issue in contract workflows: automation can move a document, but it should not silently decide who may accept changed liability. Kweku’s experience with a changed clause is a useful example of why removing manual steps requires clearer ownership, not less of it.
Measure the work that disappeared
If “jobs created” is the only question, founders learn to present headcount as the outcome. That encourages a strange kind of theatre. The company receives more credit for hiring someone to manage a broken process than for fixing the process before the hire becomes necessary.
A better assessment would ask what changed operationally.
What work no longer needs to be performed by hand? Which decisions still require a person? How much recurring capacity did the company gain? What new failure modes did the automation introduce? Where does an exception go when the normal path breaks?
Those questions do not assume that automation is good. They make the founder prove that the change works.
The same discipline applies before a customer handoff. A workflow is only useful when the next person can see what happened, what remains unresolved and who owns the decision. That is why a reliable post-signature handoff matters more than a demonstration that completes the happy path.
For an awards application, I would still answer the field directly. If the number is zero, write zero. Then explain the decision without dressing an avoided hire up as a social impact claim.
The defensible case is narrower: the automation removed a defined body of recurring work; the company retained human review where judgment or authority mattered; and the saved commitment gave the existing business more room to survive or invest elsewhere.
What I would record before approving it
Kodak’s 1975 prototype did not arrive with a settled commercial answer. It revealed that the company’s established unit of value, film, might eventually become the wrong unit for judging a camera.
The employment field created the smaller version of that problem for me. It asked for a familiar output after the operating model had changed underneath it.
Before approving the next automation, I would record the manual process as it exists, the proposed system boundary, the exceptions that remain human, and the role we believe will no longer be required. Then I would review the decision after real usage, because an avoided hire counts for little if customers wait longer or the existing team absorbs the failures.
At 2:14 PM, the truthful number could fit in the field. The useful answer belonged in the evidence behind it.
Comments
No comments yet.