Almost every conversation about technology starts from the assumption that the answer will be to build, buy or integrate something. Treating that assumption as automatic is where projects that should never have existed begin.
The first question
Before proposing a solution, the question is not what can be built, but whether there is a sufficient reason to intervene.
When the problem is not technological
What looks like a technology problem can in fact be:
- a process problem;
- a data problem, with data captured wrongly from the start;
- an operating volume that does not yet justify automation.
Calculate before building
Before starting, estimate the return: how often the process occurs, how long it takes each time, and what it will cost to maintain what is built. We have seen organizations invest months of development in automating a process that happened five times a month and took twenty minutes each time.
When there is a reason to intervene
If intervention is justified, the scope must address the cause, not the symptom. Two examples of our approach:
- Information scattered across systems that do not talk to each other: connect the existing systems, without replacing them.
- Repetitive tasks within a decision process: automate them without touching the specialist’s decision.
The criterion
Judgment is not measured by how many projects are executed, but by how often the right recommendation — build or not build — turned out to be the one the business actually needed.