Our internal tool is live and every department wants more features. How do we stop it becoming another ERP project?
Give the application a clear operating boundary and one accountable business owner. Write down the workflow it supports, the users it serves, and what it deliberately does not manage. Evaluate new requests against that boundary before anyone starts prompting Lovable or Replit to build them.
Fast development makes additions easier to produce. It does not remove the coordination, training, maintenance, and reporting obligations that each addition creates.
Separate incoming requests into three groups:
- Defects that prevent the agreed workflow from working correctly.
- Improvements that make that same workflow easier or more reliable.
- Expansions that introduce a new process, department, or system responsibility.
Resolve defects according to operational risk. Review improvements against evidence from actual use. Treat expansions as new investment decisions, with explicit ownership and an explanation of what existing system or manual work they would replace.
Use EOS disciplines without turning every feature request into a Rock. Keep unresolved business problems on the appropriate Issues List. If a substantial expansion becomes a genuine quarterly priority, define a Rock with one owner and a clear completion condition. Routine fixes belong in the normal work queue.
Ask a useful question before approving an addition: does this simplify the company's operating system, or merely move more complexity into custom software?
For Batteries Included, the recommended approach is a focused application that earns expansion through demonstrated use. Maintain a visible list of deferred requests and explain the decisions. A deliberate boundary protects the useful tool you already have.
Category: Internal Tools