The best internal tool is often smaller than the team first imagines. Not because ambition is wrong, but because leverage hides in repetition: the same decision made weekly, the same reconciliation performed monthly, the same question answered by interrupting the same senior person.
Find the friction that recurs
The discovery method is unglamorous — watch work happen and count. Which tasks recur on a schedule? Where do people maintain private spreadsheets as insurance against a system they don’t trust? A tool that removes one repeated hour from thirty people outranks a platform that might someday remove many hours from hypothetical people.
Choose the boundary like an engineer
Narrow scope still deserves serious design. The question is where to cut so the tool stays true:
- One job, stated in a sentence. If the sentence needs “and”, it’s two tools.
- Inputs you can own. A tool fed by hand-maintained data rots with every typo.
- An output that ends a debate. The strongest tools settle recurring questions — what’s available, what changed, what’s blocked — rather than adding another place to look.
Let adoption shape version two
Small tools earn the right to grow. Usage patterns tell you which adjacent capability is actually wanted — not by roadmap speculation, but by watching what users do with the edges of what you gave them. Build that next. Repeat.
Narrow scope is not a lack of ambition when it creates room for real learning. The platform your organization eventually needs is usually discovered through three small tools that worked, not specified in advance by someone guessing.