How to Go From First Call to Kickoff in Thirteen Days

On the first Tuesday in January I had an introductory call with the chief executive of a licensed apparel company. We talked through how the work would go. Find where the time is bleeding. Ship one or two things fast. Stack from there. I told him everything we built would be his, the code included, running on his own servers.

Thirteen days later we held the kickoff. His operations lead said what she wanted from a partner in one sentence: somebody who would tell her the truth and deliver what they promised.

Thirteen days is not unusual for a company that size. It is the reason I work with them.

Where I chose to work

The company I want on the other side of the table is somewhere between two million and five hundred million dollars in revenue. I will go all the way down to two million, because I like that space.

A lot of the AI market is built for much larger companies, and for them that makes sense. They have data teams, formal procurement and the patience for a long program. The companies I work with mostly have none of those. That turns out to be an advantage, because the things that stretch a program out are things a company this size never had to build.

Three parts to that.

The person with the problem is in the room

In a company this size, the person who does the work and the person who can change it are a hallway apart. Often less.

When I ask who keeps the numbers that go into the Monday meeting, the chief executive knows the name. When I sit with that person and they tell me how long something takes, the executive can hear it that same week, from them, not from a slide three layers later. The first step of our method is Find the Bleed, and in a small company the bleed has a face and a desk.

That matters more for AI than for most projects. A good build is designed with the person who does the job, tested by them and corrected by them. If they sit three reporting lines away from the decision, every correction is a meeting. If they sit next door, it is a conversation.

A decision takes a week, not a quarter

Most of what I need from a client is a yes or a no. Build this first. Not that. Give us access to this system. Let two people test it on Friday.

At a company in this range, those calls get made by the people who own the outcome, and they get made quickly. The person signing is usually the person who will feel the result.

I have seen this from the inside. The clearest payback I ever got came from automating invoicing and collections at a services company I ran. It went quickly because the person who felt the pain and the person who decided were the same person, and it was me. There was nobody to convince.

The same shape shows up in small companies constantly. The loop from "this is wrong" to "fixed" is short, so the team learns what AI is good for by using it, in weeks.

A small team can change the whole company

The part people underestimate is reach.

In a company of a hundred people, one system for the sales team touches a real share of the company. One reporting layer for the chief executive changes what every department gets asked for on Monday. One person deciding what to build and a few builders can see across the whole business at once, because the whole business fits inside one discovery.

That is also why I bring a team rather than advice. A company this size has no spare engineers. A plan that needs the client to find three of them is a plan that waits. The person deciding what to build should sit in the same meetings as the people building it, and the client should not have to hire any of them to get started.

What the client does need to supply is an owner. One person who cares about each system and will tell us, the same week, when it is wrong. Most companies this size already have that person. They are usually the one who set up the first call.

If you run a company this size

You already have the head start. The way to lose it is to borrow a process designed for a company ten times your size: a committee, a long evaluation, a strategy document before anything ships.

Try this instead. Write down the one piece of work that bothers the person who signs the checks the most. Next to it, write the name of the person who does that work. Then put the two of them in a room for half an hour this week and have the second one walk through it on their screen, start to finish, on a real example.

If you can do that by Friday, you can decide what to build the week after. That is the whole advantage, and it is sitting in your building already.