Operational success is about much more than keeping up with the latest AI model. Today, I believe that much of that success depends on what the organisation has been doing all along.

By Tiago Veríssimo, CEO at WYnova

Seven months ago, I was invited to lead WYnova, a company dedicated to creating and implementing AI solutions. My first reaction was: “But what do I know about engineering?” It was the biggest bout of imposter syndrome I have ever experienced.

Let me give you some context. I consider myself a 4×4: an unapologetic generalist, a biologist by training who became a marketer, an entrepreneur at heart and naturally curious — but anything but an engineer.

Because I have a certain tendency to jump into the deep end, I accepted. Seven months later, after dozens of meetings and proposals, I realised that the role was never really about an individual’s technical ability. It was about being able to read organisations and understand their ecosystems, anticipate where the world is heading, know which people to surround yourself with, and define the strategy needed to keep the company relevant.

Operational success is about much more than keeping up with the latest AI model. Today, I believe that much of that success depends on what the organisation has been doing all along: its foundations, infrastructure and databases; how its internal and external processes are structured; and how well its teams are equipped.

The word that best defines all of this is resilience. And there is nothing more poetic, to me, than explaining why resilience — both within a company and across an ecosystem — is critical to the success of both.

An organisation is a balanced ecosystem in which its different parts have found a way to function together in pursuit of survival and growth. It is not perfect, but it finds its balance: not every employee is highly efficient, not every process is necessary, not every department works well, not every leader is good, and not every hierarchy is aligned with reality. Over time, things adapt. Processes change, people come and go, departments are slightly restructured, and leaders move on.

In ecology, this ability to absorb disturbances and reorganise without losing function has a name — and a date: resilience, defined by C.S. Holling in 1973 in his seminal paper Resilience and Stability of Ecological Systems. When faced with a shock, a resilient ecosystem withstands it and finds a new equilibrium. A fragile ecosystem collapses.

Now, suddenly, you introduce a catalyst — a mutation, a disruption — that you claim will be beneficial, help everyone and deliver on the KPIs. That is wishful thinking. It may happen, but certainly not immediately. What happens first is something else:

• A new lens for evaluating people emerges. The distinction is no longer simply between those who are efficient and those who are not, but between those who have the intellectual flexibility to integrate new tools and those who do not.
• Processes genuinely change, disrupting routines and individual efficiency. Instead of following the new process, many people will find alternative routes because those are where they feel comfortable — and the system as a whole stops working.
• Departments that were already struggling can suddenly make catastrophic mistakes.
• Leaders who were not particularly good become even less effective. Hierarchies shift, new positions emerge, egos get bruised and power games begin.

This last point deserves a name of its own. In biology, the position of each species in the food chain is called its trophic level. When the environment changes, some species move up the chain while others simply disappear. The exact same thing happens in companies. A junior analyst who masters the tools suddenly starts feeding insights into management decisions. A middle manager whose value once came from controlling the flow of information sees that information begin to circulate without them.

None of this appears on the organisational chart, but you can feel it. And while everyone involved may be perfectly well-intentioned, this is where bruised egos and power games are born.

Stanford’s Digital Economy Lab studied 51 successful AI implementations in companies as part of its Enterprise AI Playbook and found precisely these situations occurring: organisations implementing the exact same solution with drastically different results. In the cases studied, 77% of the most difficult problems were not technical. They were related to change management, data quality and process redesign. And among the implementations that ultimately succeeded, 61% had gone through at least one failed attempt.

So, if we define resilience in a business context as the ability to absorb change, then comparing the technology implemented by two companies is not where our attention should lie. Instead, we should be looking at what happened inside the organisation when change was introduced.

A good example of this in practice comes from 1986, when Stephen Barley published a study on the arrival of CT scanners in two radiology departments that developed completely different structures around the same machine. The title says it all: Technology as an Occasion for Structuring. Technology is an opportunity to restructure. And with AI, that opportunity is even greater, because we can rethink the way an organisation’s intelligence is communicated, managed and shared.

The problem is that natural selection will do its job regardless, and it is worth remembering how it works. Yes, it eliminates what is inefficient. But when unviable mutations accumulate, it stops pruning them one by one and takes down an entire ecosystem instead.

A company with poorly designed processes, AI implementations that cost more than the return they generate, and teams working against their tools will not die because of one single problem. It will die because of the combination of all of them.

Companies that have been doing the unglamorous groundwork — strengthening their foundations and investing in their teams — have been building resilience. And they will create an increasingly wider gap between themselves and those who remained focused on quick and easy fixes.

Of course, there are quick wins that can be implemented, and even the best plan cannot mitigate every impact. But there are three things organisations can do to build resilience before it is put to the test:

• Budget for the first failed attempt as a normal cost of the process, because Stanford’s 61% shows that failure is part of the journey for those who ultimately succeed.
• Decide upfront which processes need to be redesigned, rather than automating a bad process and simply producing mistakes faster.
• Pay attention to who starts evaluating whom once the tool is introduced, because a new evaluation framework will emerge whether leadership intervenes or not.

And so we return to the beginning: AI is not just for engineers; it is also for generalists who can understand the whole by looking at its parts.

Who would have thought?

In fact, what would my genetics professor say today? She saw me walk into her lab for the first time wearing flip-flops and a lab coat and predicted: “You’re going into marine biology, aren’t you?”

No, professor. I’m going into AI — and no amount of imposter syndrome is going to stop me.

Opinion article by Tiago Veríssimo, CEO at WYnova, originally published in Digest Executive.