Most AI advice is written for companies with a change team, a data warehouse, and a year of runway. A company of forty people has none of those. It has two or three chores that eat somebody’s week, a stack of ordinary software, and no patience for a project that pays off eventually.
So the work looks different at this size. One process, one system, built inside the tools you already pay for, doing the job whether or not anyone is thinking about it.
Client performance reports written, published, and rendered without a person assembling them.
01
The first system has to pay for itself in hours somebody can already name.
Picking the right first job matters more than picking the right technology. It should be repetitive, it should follow rules most of the time, and somebody should be able to say how many hours a week it currently takes. If nobody can say that, it is not the first job.
Everything else waits. A second system is a decision you make once the first one has been running long enough to trust, on evidence rather than appetite.
02
What we sell was built under the constraints you have.
Our own companies are small businesses: a marketing agency, a distributor, the group around them. Nothing we run needs a dedicated administrator or an expensive platform underneath it, because none of us had one to spare when it was built.
The reporting system is the plainest example. It writes and publishes the monthly client reports at our agency, work that used to land on the account managers by hand, every month, whatever else was happening.
03
You start with the same fixed-fee audit as everyone else, at a price agreed before it begins. At the end you get a written verdict, a build price, and the case against building at all. Small companies hear do not build more often than large ones, because the volume behind a chore has to justify the work, and the start of the engagement is the right time to say so.
Proof
Client performance reports written, published, and rendered without a person assembling them.
Runs every month at Keller Creative, our marketing agency.
It was built for a small agency under the same constraints you have, and it has written and published the monthly client reports there ever since.
Scope
The chore eating the most hours, done properly before anything else.
Built inside the tools you already pay for, not on a new platform.
A price agreed up front and a written verdict at the end of it.
One process built and put into service, not a pilot to evaluate later.
Systems that run without anyone hired to mind them.
Documentation your team can read, and an owner who checks the output.
FAQ
Working out which one or two jobs in a small company are worth handing to software, building that, and being straight about the rest. At this size the deliverable is a working system, not a strategy deck.
Ten people is a fine place to start if the chore is real and repeats. Below that the honest answer is often an off-the-shelf tool, and we will tell you which one.
A fixed audit fee, agreed before we start. It is never credited toward a later build, because a verdict you can trust has to cost the same whichever way it goes. If we tell you not to build, you keep the written verdict and the spec.
No. Every system is built to run without a caretaker, inside software you already use. Someone on your side owns it, but owning it means reading what it flags, not maintaining it.
How We Start
AI enablement that ends in daily use: systems built into Slack, email, and the tools your team already opens, plus the defaults, training, and ownership.
automation consultingAutomation consulting that starts with a fixed-fee audit and ends in a written verdict, including do not build. Advice from a team that runs what it sells.
One fixed fee. A build-or-don’t-build verdict in writing.