Every organisation has a WDNW system. Most have several.

A process that runs quietly in the background, consuming time and money, that nobody questions because it has never visibly broken. The moment someone asks what it actually does, the honest answer is usually some version of “we are not entirely sure, but we are afraid to touch it.”

Startups are not immune to this. Technical debt accumulates faster than most founders realise. A workflow stitched together in year one, a third-party integration nobody remembers choosing, a report generated every Monday that nobody reads but everyone continues to produce. These are not dramatic failures. They are quiet drains.

This is one area where I think AI genuinely earns the word superpower. Not in building new things, but in helping us see what already exists with fresh eyes. AI tools can now map workflows, surface redundancies, flag processes with unclear ownership, and help teams ask the question they were too busy or too hesitant to ask themselves: why are we still doing this?

The best technical debt reviews are not always led by consultants or new CTOs brought in to shake things up. Sometimes all it takes is a founder sitting down with an AI tool and asking honest questions about their own workflows. No agenda, no politics, just a clear audit of what exists and why.

And you do not need a fancy specialised auditing platform that costs a fortune to get started. Tools like Claude Code or GitHub Copilot are more than capable of helping you take those first steps. Start small, pick one corner of your codebase or one workflow, and go from there. Do not try to churn the ocean in one go. With frontier AI tools, there are no real experts yet. Everyone is figuring their way through. The important thing is to keep moving, keep experimenting, and keep learning, carefully and with guardrails in place.

How many WDNW systems are hiding in your organisation right now? If you are not sure where to begin, I am happy to think it through with you.