The honest answer is usually buy. Here is the short checklist for the times it genuinely is not.
Every internal-tools conversation attracts strong opinions. Strip out the ideology and it comes down to a few questions.
Buy when
- A mature product covers 80% of the need out of the box.
- The process you are supporting is common across many companies (expenses, ticketing, scheduling, CRM).
- You do not have the capacity to maintain software for years, because that is the real cost, not the build.
- The vendor's roadmap is heading where you are heading anyway.
Build when
- The process is a genuine differentiator or is unusual enough that products fight you.
- Integration with your own systems is the hard part, and you would be building that glue regardless.
- The data is too sensitive or too central to hand to a category of SaaS you do not fully trust.
- You have, and will keep, the capacity to own it. Be honest here.
The middle path
Often the right answer is buy the platform, build the thin layer. A low-code platform or a well-chosen framework for the 20% that is yours, sitting on bought infrastructure for the 80% that is not. That keeps the maintenance burden proportionate to the value you are actually adding.
The cost people forget
A built tool needs dependency updates, security patches, a person who understands it, and a plan for when that person leaves. Budget for five years of that before you decide building is cheaper.