What the first two weeks of an engagement actually look like
Aug 5, 2026 · Engagement · 2 min read · GJH Inc.
Most of what gets written about consultancy is about the deliverable. The part buyers actually have to decide on — what the beginning of an engagement looks like — is usually vague.
Here is the beginning of ours, in the sequence it happens.
Days 01–05: going where the work is
We sit with the people doing the work and watch it happen. Not interview them about it, watch them do it. Two or three workflows get shadowed, six practitioners get asked the questions that only come up when you are watching someone do the thing for the second time.
The deliverable of this week is not slides. It is a working model of the work: which systems a task touches, which steps are loops, what a person actually has to hold in their head to get it right.
Days 04–08: what the data really looks like
In parallel, we profile the source systems. Row counts, freshness, null rates, who actually owns each table. The point is not to audit you. It is that most AI failures are data failures wearing a different hat, and you cannot know which kind you have until someone checks the data rather than assuming it.
This is usually where the interesting answer starts to appear — and it is often not the answer that was in the brief.
Days 09–14: the ranked list, with the losers named
The candidate workflows get scored against two questions: how much of the task is genuinely language-shaped, and how much is plumbing that needs cleaning before anything else. The output is a short list, ranked by expected effort and failure mode — and a list of things we explicitly recommend not building, with the reason each one is a trap.
Clients are sometimes surprised the recommendation says do less. That is the point. An honest assessment that rules three things out is worth more than a roadmap that says yes to all of them.
Day 15: your call
The recommendation memo and reference architecture land in writing. You decide: proceed, defer, or stop. No obligation, no sunk-cost pressure, and no reason to keep paying us to keep the conversation going.
The whole thing is scoped and fixed-price up front, and it is deliberately short. If we are not the right fit, better to find that out in three weeks than three months.