The Workforce Issue · Field Notes Vol. 01
The Potential Energy Problem.
Every business in America is sitting on a tank of unspent capability. The next era of workforce management isn't about replacing people with AI — it's about teaching humans and AI how to work as teammates. And the smaller the company, the bigger the opportunity.
There's a moment most mornings when I sit down at my desk, open two AI platforms side by side, take the first sip of coffee, and feel something I can only describe as potential energy. Nothing has happened yet. No work has been done. But the room is humming with what could be.
Five years ago, that same morning looked completely different. A spreadsheet. A legal pad. Fourteen browser tabs. A nagging sense that I was about to spend the next three hours doing the kind of work that isn't actually work — it's the assembly required before the work can begin.
That feeling — the gap between what a small team could theoretically accomplish and what the day actually delivers — is the most underestimated business asset in America right now. It's sitting in every branch office, every restaurant back-of-house, every city hall, every dispatcher's desk, every spa front counter. It's the potential energy of small and medium-sized businesses, and almost nobody is harnessing it.
"The gap between what a small team could theoretically accomplish and what the day actually delivers is the most underestimated asset in American business."
I — The Altitude Problem
The next era is about teammates, not tools.
Most of the AI conversation right now is happening at the wrong altitude. The headlines are about replacement, automation, headcount reduction. The reality on the ground — at least in the kinds of operations I've actually worked inside — is much more interesting and much more human.
The next era of workforce management is about securing the relationship between human and AI teammates. Not "deploying AI." Not "rolling out tools." A relationship. Which means trust, role clarity, mutual reliability, and a shared understanding of who owns what when something goes sideways.
And here's the part that gets lost: this is a workforce management problem before it's a technology problem. The same disciplines that make a good schedule, a fair territory, a functioning team — those are the disciplines that determine whether AI integration succeeds or fails. The companies that already think carefully about how their people work together will adopt AI well. The ones that don't, won't.
"The next era of workforce management is about securing the relationship between human and AI teammates — not deploying tools, not cutting headcount. A relationship."
— On the work ahead
II — Where The Real Story Is
Why small and mid-size businesses have the most to gain.
Enterprise gets the press. But the real story is in the 20-person home health branch, the three-location restaurant group, the municipal parks department, the regional HVAC company with 60 trucks. These businesses run on tight margins, lean teams, and the personal heroics of two or three people who hold the whole operation together in their heads.
Those people are exhausted. And they're holding more institutional knowledge than any system has ever captured. When the scheduler at a home health agency finally takes a vacation, the wheels don't just wobble — they come off, because nobody else knows that Mrs. Patterson on Route 4 needs her PT appointments before noon or she won't open the door.
That's where AI as a teammate changes the math. Not as a replacement for the scheduler, but as a partner that absorbs the repetitive cognitive load — the geography, the rules, the paperwork, the first-draft thinking — so the human can focus on the judgment calls only they can make. Cost is no longer the barrier it was three years ago. A small business can now afford the kind of analytical horsepower that used to require a six-figure software contract and a consultant on retainer.
The barrier now is adoption. And adoption is not a software problem.
III — Field Observations
The same pattern, in five different uniforms.
I've spent years working alongside frontline operators in environments that, from the outside, look nothing alike: home health agencies, municipal governments, hotels and restaurants and spas, large home service operations, sales teams trying to figure out territory and quota at the same time. Different uniforms, different vocabularies, completely different problems on paper.
But underneath, the pattern is identical. The same kind of person is doing the same kind of invisible work — the work between the work — and it's that work that AI is uniquely good at picking up.
Home Health
The Branch Manager
Holds the geography, the clinicians, the patient mix, and the reimbursement model in one head. AI as a teammate means the territory map, the PTO calendar, and the coverage plan are no longer "in the manager's brain" — they're shared, queryable, and survive the manager's day off.
Municipal Government
The Department Lead
Public sector teams have decades of process and almost no tooling to match. AI doesn't replace civil service judgment — it absorbs the form-fillings, the routing logic, the cross-department lookups, and gives a department head time to actually lead.
Hospitality
The GM at 4:55 PM
Restaurants, hotels, spas — the daily prep, the schedule reshuffle, the call-out at the worst possible moment. AI teammates flatten the prep curve so that when service starts, the human is doing what humans do best: reading the room.
Sales Teams
The Quarterly Planning Meeting
Territory design, account assignment, pipeline review — the spreadsheets that nobody enjoys building and everybody is forced to argue over. AI partners turn those debates into data conversations and free reps to do the thing they were hired for: talk to customers.
Home Services
The Dispatch Whiteboard
HVAC, plumbing, electrical — the routing puzzle is constant and unforgiving. AI teammates handle the optimization, the rebalancing, the "what if the 2 o'clock cancels" thinking, so the dispatcher stops solving puzzles and starts managing the day.
The Common Thread
Invisible Work
Every one of these roles spends the majority of their time on work that never shows up in a metric — the thinking, the prep, the coordination. That invisible labor is exactly what AI does best. The visible work, the human work, finally gets the room it deserves.
IV — The Only Problem Left
The adoption problem is the only problem.
Here is the part that small and mid-size operators need to hear, because nobody is saying it clearly enough: introducing AI to your workforce is possible right now, on a small business budget — but it has to be done in small steps to ensure your people actually adopt it.
I have watched companies spend serious money on capable software and get nothing back, because they tried to launch it to the entire team on a Monday with a 90-minute training and a thumbs-up emoji in the company Slack. Three weeks later, half the team is back to spreadsheets and the other half is quietly resentful. The tool didn't fail. The rollout did.
The pattern that works — the one I've watched succeed across home health, hospitality, municipal, sales, and home services — looks something like this:
Step 01
Start with one painful, repetitive task
The schedule, the report, the routing decision, the first-draft email. Pick the task that generates the most private groaning on your team. The one that the best person in the role has to sit down and do anyway, even though it's beneath what they were hired for.
If you pick well, the first win is self-evident. No change management deck required.
Not a vision. Not a transformation. One task that one person hates doing every week. Solve that one thing well and let it speak for itself.
Step 02
Pick the right first teammate
Every team has someone who is curious, respected, and not afraid of being early. That's your first AI partner — not the loudest skeptic and not the most senior person.
When a respected peer starts quietly getting more done with less friction, the rest of the team notices. That's the adoption curve you want — pulled from inside, not pushed from the top.
Curiosity travels through a workforce faster than mandates do.
Step 03
Make the human the editor, not the operator
The AI drafts. The human edits, approves, and ships. That ratio matters. It builds trust quickly because the human remains the accountable party, and it preserves the judgment of the people who actually understand the work.
Flip that ratio too early and you lose both the quality and the people.
Nothing goes out the door that the human didn't sign off on. That ratio is what builds trust.
Step 04
Measure time returned, not technology adopted
Forget dashboards of "AI usage." Ask one question: what did you get back this week that you didn't have before? Ninety minutes on a Friday afternoon. A quieter Monday morning. The ability to take a real lunch.
That is the metric that matters to the people you're asking to change. Everything else is theater.
The metric that matters: what did you get back this week that you didn't have before?
Step 05
Expand only when the team asks for it
Not before. An expansion driven by a leadership decision will stall. An expansion driven by a colleague saying "wait, how did you do that?" builds its own momentum.
Trust the pull. Ignore the pressure to scale on someone else's timeline.
The pull from the team is the signal. When the second person asks how to do what the first person is doing, you've earned the right to roll it out wider.
V — What It Really Means
What securing the relationship actually looks like.
The phrase "securing the relationship between humans and AI teammates" can sound abstract until you've watched it succeed or fail in a real operation. Up close, it comes down to a few unglamorous things.
It means role clarity: who decides, who drafts, who approves, who is accountable when something goes wrong. It means boundaries: this AI handles scheduling drafts; it does not handle clinical judgment. It does not handle hiring. It does not handle the conversation with the family. It means visibility: the team sees what the AI saw, what it decided, and why — no black box, no surprise outputs.
And most importantly, it means protecting the human dignity of the work. The frontline worker doesn't want to be told a robot is going to make their job easier. They want to be shown that their judgment is the thing the company is trying to free up — that the AI is there to handle the part of the job that was never the point of the job.
"The frontline worker doesn't want to be told a robot will make their job easier. They want to be shown that their judgment is
the thing the company is trying to free up.
"
— On dignity at work
Every small and mid-size business in this country has a tank of potential energy sitting in its operations right now. It's in the schedule that takes nine hours to build. The territory map that hasn't been redrawn in three years. The Friday afternoon report that everyone dreads. The dispatcher's whiteboard. The GM's clipboard. The branch manager's brain.
The next era is the era we finally figure out how to
bottle that energy
— not by replacing the people who built it, but by giving them a teammate who can help them carry the load.