Buy off-the-shelf when the problem is common and the tools already exist. Build custom only when your workflow, your data, or your compliance needs make the off-the-shelf answer break.
Most businesses need the first, not the second. A vendor who reaches for a custom build on every problem is selling hours, not solving yours.
Default to off-the-shelf
Zapier, Make, and existing SaaS solve the majority of automation problems out of the box. They are cheaper, faster to deploy, and maintained by someone else. If a configured tool covers 80% of what you need, that is usually the right answer, not a reason to build from scratch.
When custom is worth it
- Your workflow is genuinely unusual and no tool models it.
- You have proprietary data an agent needs to reason over (RAG).
- Compliance or data residency rules mean the data cannot leave your control.
- Volume is high enough that per-task pricing on an off-the-shelf tool costs more than a build.
- The capability is a real competitive edge, not a commodity.
The expensive mistake, both directions
Building what you could have bought burns months and money on a system someone else already maintains for a subscription. Forcing an off-the-shelf tool onto a workflow it does not fit creates a fragile mess held together with workarounds. Both are common. Both are avoidable.
A cheap way to decide
Map the workflow first, then check whether a configured tool covers most of it. If it does, buy and configure. If it does not, build only the part that is missing, not the whole thing. An Operational Audit does exactly this mapping, which is why it is the honest first step before anyone quotes you a build.