The first mistake is starting from the tools. New CRM, faster internet, a reporting platform, something with AI in it. That is starting from the answer, and it produces a plan that cannot survive a changed priority because nothing in it explains why any item is there.
We start from what the business is trying to do in the next eighteen months, and work backwards to the technology that either enables it or quietly prevents it.
What we actually do
Understand the operating reality — not the org chart, but how work actually moves, where it stalls, and which two people everything depends on. Find what is holding you back, which is frequently not the thing anyone complained about. Sequence the work, so each phase is useful on its own rather than a down payment on a benefit that arrives in year three. And put numbers on it, including the running costs that quotes tend to leave out.
Honest about horizons
Three-year technology plans that name vendors and go-live months are fiction with a Gantt chart. What holds at three years is direction — "our systems should talk to each other rather than be re-keyed" — while specifics belong in the next two quarters, where you can actually be accountable for them.
So a roadmap from us is deliberately uneven: precise about the next six months, directional after that, and explicit about which assumptions would change it. That is the shape we argue for in IT planning that survives the year.
Including what not to do
A strategy that only adds is a budget, not a strategy. Ours name the things worth stopping: the subscriptions nobody uses, the server that should not still be in the office, the custom system that costs more to maintain than to replace, the integration that was a good idea for a workflow you abandoned.
We are also willing to say that the answer is no new technology this year. That advice has cost us work and we would give it again, for the reason set out in why a partner beats the cheapest quote.
What you walk away with
A written roadmap with sequencing and costs. A short list of things to stop. The risks you are currently carrying without having chosen to, named plainly. And enough reasoning that you can defend it to a board, or change it in six months without starting again.
Where AI fits in
Increasingly this conversation includes AI, and the useful version of it is unglamorous: which specific tasks, what does the data need to look like first, what does it cost to run, and who checks the output. A business whose systems do not talk to each other is not ready to layer AI over them, whatever the demos suggest. We cover the sequencing in building an AI-ready IT strategy.
