In this guide
The wrong place to start
Software projects often begin with a feature list: a new portal, a dashboard, an integration, an AI assistant. Starting with a feature list can feel like progress before the underlying problem is understood.
But a feature list is a guess about the answer. If the guess is made before anyone understands how the work actually runs, the software can be well built and still miss.
In short
Understand how the organisation works first: how the work moves, where context gets lost, who decides and what the organisation is trying to achieve. Only then choose what to build.
What to understand first
Before deciding what technology should do, it helps to answer four questions:
- How does the work move? Who starts it, who touches it, where it waits and where it ends.
- Where does context get lost? The handovers, side spreadsheets and email threads where information drops out of the system.
- What does the team need to decide? Which decisions need judgement, which are routine and who has the authority for each.
- What is the organisation trying to achieve? Not just this quarter's fix, but where it is trying to position itself.
None of these are technical questions. All of them change what the right technology looks like.
The fastest way to build the wrong software is to begin with the software.
Then the right answer becomes visible
With that picture in place, you can compare the options against the organisation's needs. Sometimes the answer is a connected system, so information stops being carried by hand between programs. Sometimes it is an AI agent taking on defined preparation work within clear limits. Sometimes it is a different way of bringing people and information together. And sometimes it is not software at all, but a change in responsibility or process.
If you are weighing up agentic software, our plain-language guide explains what it is and when it fits.
How we work
This is how we start every engagement: with the organisation, its people, its work and what it is trying to achieve. The technology comes after, and it is chosen to fit. If you would like to talk through a piece of work, start a conversation.
Questions
Frequently asked.
It usually saves time. Building the wrong thing quickly is slower than building the right thing once. The understanding stage is where the expensive mistakes are cheapest to fix.
That is a useful starting point. We still look at how the work moves today, because it often shows whether the request will solve the problem behind it, or whether something simpler would.
No. Sometimes the right answer is a connected system, an AI agent taking on defined work, or a change to how people and information come together. Sometimes it is a process change, with no new software at all.
Written by
Mare Pomana
Plumb Digital