The Decision Company

Most companies describe themselves by what they sell: consulting, software, manufacturing, logistics. Those are industries. They don't explain how a company works.

We start from a different question. What is a business? Our answer: a business is a system for making decisions. The products, the factories, the software, and the people all exist to support them. When the quality of important decisions improves consistently, the business improves. When it deteriorates, more technology rarely fixes it.

Most decisions are ordinary

Every day, people in your company decide what to buy, make, schedule, quote, approve, repair, and ship. Should we accept this order? Can we promise this date? Maintenance now, or next week? Is demand real, or are we building inventory?

No single one of these transforms anything. Their accumulated quality determines margin, cost, service, risk, and capacity. That's why we treat the decision — not the software, the model, or the report — as the unit of improvement.

Technology is a means

An ERP records transactions. A dashboard presents information. A model predicts, recommends, and sometimes acts. These systems can make a well-run operation faster and more consistent. They can also make a confused one more complicated.

So the first question is never whether AI can do something. It's whether improving this decision would materially improve the business. Only after that gets answered do we pick the method — and the right method is the simplest one that moves the number.

Data becomes information when it reduces uncertainty

Data isn't valuable because it's been collected, centralized, or put on a dashboard. It becomes information when it helps someone tell two courses of action apart before the decision has to be made — which means it has to be relevant, dependable, and there in time.

A company can own years of history and still make its important calls from a spreadsheet and a gut feeling. The problem usually isn't a shortage of data. It's turning data into evidence at the moment of choice.

Judgment is the rarest asset

Your most experienced people know things the systems don't. Which measurement is wrong. Which supplier promise is credible. Which machine sound matters. When the standard procedure should be ignored.

That judgment is difficult to replace and easy to lose — it walks out the door in a retirement letter, and none of it is written down. Before automating work, capture how those people decide: the rules, the exceptions, the tells. AI makes that knowledge easier to keep and to share than it has ever been. It's a terrible excuse to ignore the people who hold it.

Complexity is debt

Complexity accumulates one reasonable exception at a time. An acquisition brings another system. A big customer gets a special rule. A team builds a spreadsheet because the official path is too slow. Each choice made sense at the time.

Eventually the business spends more effort coordinating itself than improving. The right response isn't always more integration or automation. Some of the highest-return work is removal — an approval, a report, a duplicate record, a reconciliation that a better source of truth makes unnecessary.

Improvement is a discipline

Observe the work. Measure the current result. Find the constraint. Make the smallest change that could move the number. Measure again.

That's the whole method. It doesn't need grand language — it needs an honest baseline and the discipline to learn from what happened. The work advances only when the evidence does.

The standard

Every recommendation we make has to pass four questions:

1. Does it improve an important decision?
2. Does it simplify the business?
3. Does it preserve useful knowledge?
4. Will the company still benefit after we leave?

If the answer to any of them is no, we don't recommend it.

Better-run businesses build stronger industries. The technology is the means. Better decisions, every day, for decades, are the end.