Skip to content
Kosek Advisory Group

How we work

Enterprise rigor, adapted to the realities of running a business you own.

Four commitments govern every engagement, and they hold whether the work is a three-week audit or a standing retainer:

  1. We embed, we do not hand off decks. Most strategy work ends with a slide deck and a handshake. Ours starts there. We sit in the work alongside you and your team, so decisions get made and executed, not filed.
  2. Enterprise frameworks, right-sized. The methods we use were proven in large organizations. We adapt them to the constraints of a business you run yourself: limited bandwidth, real cash-flow pressure, and a team that wears many hats.
  3. Fixed scope, and a definition of done you agreed to. Every engagement names its deliverables in writing before it starts, so finishing is a matter of record rather than a matter of opinion.
  4. You keep the systems. When the engagement ends, you keep the operating systems, the documented decisions, and the evidence behind them. The goal is a business that runs better without us, not one that depends on us.

The standard

How a finding becomes a recommendation

Every recommendation we hand you can be walked backward to the thing that was actually observed. The chain has four links, and each one gets written down before the next is allowed to start:

1234observationcheckfindingrecommendationAI surfaces disagreements and draftsit decides nothinga named practitioner checks every findingand signs the recommendation
Artificial intelligence (AI) runs through the whole engagement, not only the two links the inner bracket marks. Where the line between the tooling and the practitioner falls, in both directions, is set out further down this page.
  1. 1. The observation. What a person said in an interview, or what a document or a system showed. Recorded with its source and its date, in the terms it was found in, not as a summary of what it probably meant.
  2. 2. The check. The observation held against something independent of it: a second interview, the system of record, the numbers. What survives becomes evidence. What does not survive is written down as unresolved rather than quietly dropped, because the thing two people disagree about is usually where the real friction lives.
  3. 3. The finding. A statement about your business that the evidence supports, carrying how confident we are in it. Where the evidence is thin, the finding says it is thin. A confident-sounding finding built on one comment from one person is the failure this step exists to catch.
  4. 4. The recommendation. The action the findings point to, alongside the alternatives we considered and set aside, and why. You get the road not taken as well as the road we recommend, so the reasoning is yours to challenge rather than ours to protect.

The reason to work this way is not tidiness: a recommendation you cannot audit is a recommendation you have to take on faith, and the calls this work touches are too expensive to take on faith. A named, credentialed practitioner is accountable for that chain. The same standard applies whether the question is a hiring decision, a vendor choice, or a strategic pivot.

The tooling

Where artificial intelligence sits in the work

We use artificial intelligence (AI) throughout our engagements, and we say so before you ask. Two questions follow from that, and both deserve a straight answer.

What it does here
Reads across interview transcripts, documents, and system exports faster and more evenly than a person reads them. Surfaces the places where two of those sources disagree, which is where the checking in step two starts. Drafts, so that more of an engagement goes to judgment than to typing.
What it does not do
It does not decide, and it does not sign. Every finding is checked by a person against its original source before it earns the word. The recommendation, the confidence attached to it, and the accountability for both belong to a named practitioner with measurement-science standing. Where a call touches a person’s job, pay, or hiring, a human makes it and can explain it afterward, on the record.

Many business owners have not brought AI into their own operations yet. An engagement doubles as a working introduction to it, applied to your real problems instead of a demonstration, and you get to watch both halves: where the tooling earns its place, and where we keep it deliberately out of the room.

The first step

How an engagement starts

Most of the time it starts with a conversation. Sometimes it starts with something already built. Both are real doors, both lead to the same work, and neither one commits you to anything.

You reach out
A free consultation first, where we work out together whether we can genuinely help. If we can, the Strategic Audit is the first step you pay for: three weeks, time-boxed, ending in a written current state, the real options with one recommendation, a 90-day plan, and a baseline you can measure against later. All four are yours to keep either way. Implementation follows only if the audit points at work worth doing.
We reach out, with work in hand
Now and then we read a business from the outside, using only what is public, and build a piece of the work before we have spoken: a site rebuilt on their own content, a pricing tool, a process laid out end to end. It arrives as a private link with nothing attached to it and no obligation in either direction. A conversation goes further when there is something real on the table to react to.

The second door changes where a conversation starts and nothing else. The consultation and the audit sit ahead of the work either way, the terms are the same, and a preview built before we met stays a preview until you decide otherwise.

Let's talk about your next stage.

Book a consultation (opens in a new tab)