There’s an insightful quote from Upton Sinclair: “It’s difficult to get a man to understand something when his salary depends on his not understanding it.” In the interview study we’re in the midst of as I’m writing this, I’ve frequently thought of this quote. It seems that people simply don’t understand that their job, or at least tasks that are part of it, will disappear or be automated in ways that were impossible before AI.
Whether it be telephone operators, paper map makers or translators, it’s not that the ones with digital tool support are outcompeting the old-fashioned ones. It’s that the job goes away completely. All the horses that we had at the beginning of the 20th century weren’t replaced by faster ones, but by cars. Soon, it will be truck drivers, taxi drivers and all kinds of operators who will be replaced by AI-driven solutions. The same is true for numerous other jobs.
Of course, we’ll create a host of new jobs. The first roles will, to a large extent, be focused on supervising AI agents to ensure that the initial mistakes are weeded out. Over time, however, we’ll see humans switch to other roles that we can’t even imagine today.
One topic that often comes up is that the AI-enabled solution needs to be as good or better than a human-driven solution. However, that’s obviously not the case. The solution simply needs to have better economics. As an example, in modern supermarkets, self-checkout is increasingly becoming the norm. The owners of supermarkets know that this leads to higher levels of theft, as well as items that were accidentally not scanned to be taken home. However, the higher losses are, by far, outweighed by the reduced personnel cost for cashiers. It’s simply a more economical solution. The same is the case for AI-enabled solutions.
The challenge is that most of the companies we’ve interviewed are focusing on integrating AI into existing workflows. This can include supporting or automating tasks conducted by individuals as part of their job or automating steps in a workflow that weren’t feasible to automate before. However, applying AI in existing workflows and business processes isn’t the right approach. We need to start from a “zero-based thinking” position where we throw out all the preconceived notions and assumptions and design a workflow and business process from scratch.
As a starting point, it’s important to consider that workflows and business processes are typically designed around the limitations of human information processing capabilities. We break processes into steps and assign these steps to different individuals because no human can keep all the relevant information in his or her head or the amount of work is simply too much for one individual. AI agents also have limitations, but these are very different from humans (eg hallucinations and a limited context). That means that in an AI-driven organization, we design workflows and business processes around the limitations of AI agents rather than those of humans. And then we insert tasks and responsibilities for humans where AI agents fall short.
Although I’m convinced that many understand this, it tends to result in significant organizational and cultural resistance. There are at least three major driving factors: the expert fallacy, fear of obsolescence and general resistance to change.
As Shunryu Suzuki wrote, in the beginner’s mind, there are many possibilities, but in the expert’s, there are few. My favorite definition of an expert is someone who will tell you why something can’t be done. This is a key factor in most organizations faced with adopting an AI-driven mindset. Senior leaders lack the technical skills to be able to guide the organization through the adoption of AI. Instead, they give this task to experts in the organization who, rather than starting from a beginner’s mindset, bring their decades of experience and insights to the initiative and fail to reinvent the company from a “zero-based thinking” perspective.
It’s important to remember that in times of rapid change, being an expert can easily become a disadvantage as you tend to have many outdated preconceptions that are so deeply ingrained that you’re not even aware of them anymore. The only antidote to this problem that I’m aware of is unbridled curiosity and a continuous asking of “why.” We need to ensure that we’re aware of our beliefs and frequently reconfirm that these beliefs are still accurate.
The second factor causing resistance is that AI is mainly adopted to improve efficiency. Efficiency typically means that the same work can be done by fewer people. This causes people to worry about their jobs. Although I know of a few who have, for instance, valiantly helped offshore their tasks to a low-wage country, knowing that it would mean the end of their position at the company, most aren’t as altruistic and don’t put the needs of the company before their own. People fear obsolescence and are often not interested in going back to basics and reeducating themselves in a new field. Instead, we tend to believe that simply because we’re good at something, it has to be important.
The challenge for companies isn’t that they won’t be able to develop products and provide services using the old ways of working. Rather, new entrants and incumbents that aggressively adopt AI will be so much more efficient and effective that we’re outcompeted. As I mentioned earlier, it’s not that we wouldn’t be able to distribute goods using horse carriages; it’s simply that trucks are so much more competitive that it makes no sense to use horses.
The third factor is the general resistance to change. Humans are at least three times more afraid of losing something than they are concerned with gaining something. This can easily be explained through our history, where, during the tribal, hunter-gatherer phase of humankind, change brought risk and risk brought death. So, even if the consequences today are quite mundane compared to those in the tribal age, we have the same emotional wiring and hence we resist change and prefer to keep things as they were.
In organizations and even entire business ecosystems, this easily leads to a deadlock where every party forces everyone else to stay the same. And every change that will have any material impact will require everyone to agree to it, with the change being blocked if even a single party vetoes it. This is, of course, how companies and even complete business ecosystems get disrupted. However, being aware of it doesn’t necessarily make it easier to drive the necessary changes. The third challenge to becoming an AI-driven organization is organizational and cultural. We discussed the expert fallacy, the fear of obsolescence and general resistance to change. However, companies that don’t fundamentally reinvent their business processes and workflows from an AI-driven or AI-first perspective are simply going to be outcompeted by the new entrants and incumbents that do. So, it’s not really like you have a choice. To quote John Chambers: “If you’re not a disruptor, you will be disrupted.”
Want to read more like this? Sign up for my newsletter at jan@janbosch.com or follow me on janbosch.com/blog, LinkedIn (linkedin.com/in/janbosch) or X (@JanBosch).
