Automating the Mundane

A common talking point in AI discourse is that AI can take tedious tasks off people’s plates, freeing them to do more important or meaningful work. It is a value proposition that is hard to argue with. Who would voluntarily choose to do mundane work all day?

However, part of that conversation reminds me of a piece of advice I came across in a book not long ago: “If you have good knives you prefer using, it also means you have bad knives. Get rid of those.”

Automating processes that are unnecessary to begin with does not add value, even if it frees up time from the infamous “undifferentiated heavy lifting”. It often masks deeper problems.

In my experience consulting with companies across a broad range of industries, sizes, and countries, I found it common for initiatives to be framed around surface-level problems that had second-order issues underneath, which would be much more impactful to address.

In UX work, I often found it helpful to identify the “moments that matter” in a customer journey. What makes an experience worthwhile, memorable, and valuable? What breaks it? To understand this, it helps to map out existing journeys and also have a clear understanding of personas’ goals.

That same framework can be used for business processes. Mapping existing processes and stakeholder goals can unearth deeper problems, as long as you do not stop at the surface level. Thus, it is important to start with a solid understanding of operational models and overall maturity.

We often hear of AI pilots not scaling well in organizations or not producing the anticipated ROI. A question I would ask - in addition to many others - is whether the pilots address important aspects or simply delegate tasks and processes that were only marginally relevant to begin with.

“This is how it has always been done, but now we do it automatically” is not a breakthrough.