When we build a digital solution, it is easy to move fast from an idea to implementation. A good prompt can help us create code or a specification, but it does not mean that we understand the business problem. Before we build, we need to understand who has the problem, how often it happens, what they do today, and what result would really matter.
Let’s use a simple example: invoice approval. At first, the request sounds clear: “We need to automate invoice approval.”
then…
- But who submits the invoice?
- Who checks the details?
- Who is allowed to approve it?
- What happens if the amount is above a limit, if the invoice is duplicated, or if some information is missing?
- What if the approver is not available?
… and more …
These answers describe the real flow and the rules that give meaning to the solution. Without them, we may automate an incomplete process or improve one step that does not solve the main problem.
For this work, we need a way to keep the context. I think of it as a durable product context repository: an organised place where the team keeps information that can change a decision, a business rule, a priority, or a test. It does not mean documenting every conversation. It means keeping the evidence, business flows, rules, decisions, and open questions that a future team may need.
It is also important to know what type of information we have. A measurement from the system is an observed fact. “Approval takes two days” may be an estimate from a stakeholder. “Automation will remove the delays” is an assumption until we validate it. We should record the source and how sure we are about each statement. Later, we can check it again: is it still true? What evidence supports it? What have we learned that confirms or changes it?
This helps us choose what to learn first. Not all unknowns are equally important. A question should be investigated early when its answer can change what we build, or even our decision to continue.
Keeping this context becomes important when the product changes. In three or four years, the interface, database, or technology may be replaced. But the team should still understand who can approve an invoice, what must be audited, and which conditions must be met before payment. Code shows what the system does today. Business context explains what it needs to continue doing, and why.
Before we build, we should understand the problem, check what we know, and keep the reasons behind our decisions. This way, we can change the technology without losing the business logic that makes the solution valuable.
Comments
Post a Comment