The most expensive sentence in an ICT programme is “we will sort the architecture in phase two.” Phase two is where the budget goes to live, the vendor’s A-team rotates off, and the organisation discovers it has purchased a collection of screens sitting on an ungovernable estate.
Architecture-first is not ceremony. It is the decision to make the expensive mistakes on paper, in a room with the people who will inherit the system, before those mistakes are compiled into production.
What “architecture” has to mean here
In a Nigerian startup or a growing organisation, architecture is not a C4 diagram for its own sake. It is the answer to a short list of unforgiving questions. What is the system of record? Who is allowed to change it? Where does the data live, and under which law? What happens when the round closes, the CTO leaves, or the vendor’s support contract lapses?
If those answers are not written down, they will be improvised — usually by the most heroic person still in the building at 7 p.m. Heroism is not an operating model.
A week that saves a year
A proper discovery produces artefacts a founder and an auditor can both read: decision records, a security and NDPR baseline, a landscape of what already exists, and a sequence that respects hiring and cash. That package is slower than a kick-off demo. It is faster than rebuilding the thing you demoed.
Axiforge will not start a build until that package exists. It is the least fashionable posture in a market that sells velocity. It is also the only one we will sign our name to.
Related practices