The repetitive work is the clue

Before looking for the big AI idea, it is worth looking at the repetitive work people are still doing manually every week.

The work that keeps repeating

I keep coming back to this when thinking about AI products: the best place to start is not always the shiny idea. Sometimes it is the work that looks too boring to be interesting.

The same information copied from one place to another. The same update rewritten every week. The same question asked in three different places. The same person remembering the process because nobody has ever turned it into a proper system.

That is usually the clue.

Not because admin is exciting. It is not. But because repeated work often shows where something underneath has not been designed properly yet.

The quieter problems

A lot of the AI conversation starts with the big idea: the clever agent, the smart chatbot, the product that sounds impressive when you explain it.

Some of those ideas are useful. Some will become proper products. I am not against the shiny ideas. I just think they can make people miss the quieter problems.

The report someone keeps rebuilding manually. The approval that stops unless someone chases it. The customer handover that relies on memory. The knowledge that lives across a spreadsheet, three documents, and someone’s head. The update that gets rewritten because every person needs it in a slightly different format.

None of that sounds like the future. But it is real work. And it is work people are still doing every week.

That is the bit I find interesting.

The cost is not just time

The annoying part is not just that repetitive work takes time. It is that it uses up the bit of people that should probably be used for thinking.

Checking, copying, chasing, reformatting, searching for the latest version of something — it all looks small on its own. Together, it becomes a background tax on attention.

And attention is usually the thing small teams have the least spare.

Where the product thinking really sits

I have been testing a lot of AI and build tools recently, and this is the part I keep noticing. The tools are getting better quickly. Some of them are genuinely useful. But the useful bit is rarely just the output.

It is whether the output fits into the work properly.

Where does it go? Who checks it? What happens next? Does it save time, or does it create another thing to review, copy, fix, or chase?

That is where a lot of the product thinking seems to sit. Not only in what the AI can generate, but in where it belongs.

A chatbot on top of a broken workflow might help. It might not. Sometimes it just becomes another thing people have to manage.

Useful can look boring

The more useful opportunity may be much less impressive from the outside: pulling the right context together faster, turning messy information into a clean first draft, checking what changed, moving the next step forward, or reducing the number of small manual actions someone has to remember.

That does not sound especially futuristic. But it feels closer to the actual problem.

This is why I like looking for repetition. Where does the same thing happen again and again? Where does work stop unless someone remembers to move it along? Where is information being carried manually between tools, people, or documents? Where has “that is just how we do it” quietly become the system?

Those places are not always obvious opportunities at first. They can look too boring. But boring is often a better starting point than people give it credit for.

What should stay human

Not everything needs to be automated. Some work needs judgement. Some work needs care. Some work needs context. Some work needs a person who can look at the situation properly and make a call.

But when people are doing repeated, low-judgement work just because the system around them is messy, that feels worth noticing.

The repetitive work is not always the boring bit to ignore.

Sometimes it is the clue.

Linda Eriksson

I write about AI, automation, and the practical work of turning ideas into useful software.

The repetitive work is the clue

Linda Eriksson

Build software products designed for real-world use.