What simplifying actually removes

There is a version of simplification that looks productive but is not. You take a complex process, draw a cleaner diagram of it, rename the steps, and call it simplified. Nothing has been removed. The diagram is just tidier.

Real simplification removes things. Steps that are not earning their place in the process. Meetings that exist because they have always existed. Reports that someone generates and no one reads. Approvals that are nominally there for quality control but practically function as bottlenecks.

The Simplify step in the Lion Ethos is not about aesthetics. It is about subtracting what does not belong.

I work mostly with SMEs in Trøndelag, and the complexity I find in their operations is almost never intentional. It accumulates. A workaround gets built for a situation that no longer exists. A step gets added after something went wrong, and then the thing that went wrong gets fixed properly, but the extra step stays. The organisation carries the weight of its own history.

The right question to ask about any process is not "how do we make this work better?" It is "does this step need to exist at all?"

That question makes people uncomfortable. If a step exists, there is usually a person responsible for it. Asking whether it needs to exist feels personal. It is not. It is operational. But you have to separate the two deliberately, or the conversation never happens.

When I map a client's current workflow before introducing any AI, I spend most of the time in that mapping phase removing steps before I ever talk about automation. Automating a broken or unnecessary process makes it fail faster and at higher volume. Simplifying first means the system you build is doing work that actually needs to be done.

A logistics company I worked with in Frøya had a three-step internal approval process for orders above a certain value. The threshold had not been updated in six years. Inflation had moved it to the point where sixty percent of ordinary orders now required approval. The process existed. No one had questioned it because no one had mapped it recently.

Removing those two redundant steps did not require AI. It required someone asking a direct question about whether the rule still made sense.

That is what simplification costs you: the assumption that things are the way they are for a good reason. Some of them are. Plenty are not. Finding out which is which is the work.

Once you have simplified a process properly, automating or systematising it becomes straightforward. The complexity is gone. What remains is the actual operation - the steps that are there because they need to be there, running in the order they need to run.

The businesses I see struggle most with AI implementation are the ones that try to skip this step. They add AI into a complex process and the AI inherits the complexity. The output is unpredictable. The errors are hard to diagnose. The system requires more oversight than the manual process did.

Subtract first. Then build.

Murphy Alex builds operational AI systems for Norwegian SMEs from Frøya, Trøndelag. IPRESTANDA is at iprestanda.com.