The Coordination Tax Is Real. Small Businesses Pay It Differently.

Eric Schmidt knows something about coordination work.

As Google’s chief executive, he spent much of his week in meetings — by his account, about 12 hours a day, five days a week — coordinating people and putting out fires. In a recent conversation, Schmidt argued that AI platforms will reduce what he calls the “coordination tax”: the work of planning, tracking, sharing information, and keeping large organizations moving in the same direction.

His estimate was blunt. In a company of 10,000 people, he suggested there may be a couple thousand whose primary work is that coordination layer — people a large company might eventually decide it does not need.

That may be a plausible conversation inside a company with 10,000 employees. It is the wrong conversation inside a company with 25.

A small business does not have spare “glue people”

In a large company, coordination can become its own department. In an independent business, the people doing the coordinating are also doing the business.

They are the owner answering the customer question nobody else can answer. The office manager who knows which invoice went out, which delivery is late, and which employee needs a form. The operations lead who carries the schedule in his head. The salesperson who follows up because she knows the relationship matters.

Calling those people a tax misses what they actually are: the connective tissue of the business.

They cannot simply be removed. If they left, the knowledge, judgment, and customer trust would leave with them. The problem is not that they are doing coordination work. The problem is how much of their week that lower-value work consumes.

The work underneath the work

Most coordination is not one dramatic task. It is a hundred small transfers of information:

• “Has that order shipped?”
• “Can you send the quote again?”
• “Who approved this change?”
• “What did we decide in yesterday’s meeting?”
• “Has anyone followed up with the customer?”
• “Where is the latest version of that document?”

Each question requires somebody to stop, look something up, ask somebody else, write a reply, or update a tracker. The work is necessary. The repeated searching, copying, summarizing, and reminding around it often is not.

That is where an AI agent can help. Properly set up, an agent can gather the relevant information, draft the status reply, summarize the meeting, prepare the follow-up, or maintain the checklist. A person still reviews anything important before it goes out. The agent carries the routine motion; the employee keeps the judgment.

Do not start by choosing a tool

The first step is not buying another AI subscription. It is finding the coordination work that repeats often enough to map.

Pick one week and notice where information gets stuck. Which questions are asked again and again? Which updates require three people to produce one answer? Which follow-ups happen only when somebody remembers? Which meetings end without a written decision or a clear owner?

Then choose one narrow workflow. A good first candidate has four traits:

1. It happens every week.
2. The information already exists somewhere accessible.
3. The output follows a recognizable pattern.
4. A person can review the result before it affects a customer or employee.

Customer status questions, meeting summaries, quote follow-ups, and routine internal updates often meet those tests. Sensitive decisions, personnel matters, and judgment calls should stay with people.

Schmidt’s survivor list looks different at small scale

Schmidt pointed to inventors, salespeople, and leaders as the kinds of people who remain essential. In a small business, that describes much of the team. The dispatcher who solves problems creatively is an inventor. The lead who keeps a difficult customer is doing sales and leadership at once. The office manager who knows how everything connects is carrying institutional knowledge no platform can replace.

The goal is not to make those people easier to cut. It is to stop spending their limited time on work a machine can prepare for them.

For a small team, eliminating the coordination tax should mean fewer dropped follow-ups, faster answers, and more time for the work customers actually value. It should not begin with a smaller team. It should begin with a team that is better supported.

At Piedmont IT Advisors, that is the work we help businesses map: Equip the team you have with AI agents — not more hires.

First shared on LinkedIn on October 23, 2026. Read the original discussion here: https://www.linkedin.com/feed/update/urn:li:activity:7519302917494026241/

Equip the Team You Have — the training program →

Start with a free one-hour evaluation →

Related: Your Best People Are Answering the Same Questions All Day →