Learnings on AI Implementations in Supply Chain

My Experience

After leading digital transformation initiatives and businesses for over 35 years, I can say I've seen my share of the good, the bad, and the ugly when companies decided to challenge their own cultures and move to a different stage of efficiency.

As a true testament to "old dogs can learn new tricks," over the last five years, I've led dozens of AI implementations in supply chain, working with companies from multiple countries across retail, ecommerce, manufacturing, and distribution, ranging from $20M-year revenues to billion-dollar multinationals.

However, nothing I've experienced comes close to the AI transformation we are all facing today. The difference between companies that embrace AI as a business transformation and those that treat it as just another software project is dramatic, often measured in hundreds of millions of dollars in gains or losses.

A quick business case

A pharmaceutical company in the USA kept 30 weeks of safety stock for every SKU. During the implementation of their AI system, we gradually convinced them to reduce it to 20 weeks, then to 10 weeks, until finally reaching an AI-based ABC service-level model.

That change was only possible because the AI could recalibrate future sales demand and inventory levels after each new sale, and because their teams embraced the AI with a strong sponsor to guide them.

The result: a $50 million-per-year company converted $18 million in inventory back into cash flow within six months.

This series of articles shares some of the key lessons I learned while navigating cultural resistance, achieving remarkable results, and making more than a few mistakes along the way.

I hope that these insights will help your organization maximize your gains “the good” and avoid the mistakes “the bad and the ugly” while navigating this new revolution caused by AI.

Chapter 1 – The Human Component

Why AI is not a technology topic, it is a cultural one.

I understand why saying Artificial Intelligence is NOT only a technology topic can sound like an intriguing way to start a conversation. But let's stay provocative for a minute.

Yes, AI represents the peak of our technology. Building something capable of interacting, almost at a personal level, with every human being on the planet can only be achieved through breakthrough innovation and algorithms that are almost beyond our comprehension. However, all that technology still needs to be adopted by this complex, dynamic, sometimes illogical, habit-oriented creature: humans.

"Disregarding the human component in technology projects is, by far, the most common mistake companies make while leading business transformation."

If this sentence was already true in the past, it is exponentially truer now with Artificial Intelligence.

After leading hundreds of process improvement initiatives throughout my career, I have always been intrigued by how resistant humans are to change. We redesign processes, with or without technology, making people's lives easier, their jobs faster, and their daily routines less repetitive and more interesting. Yet, for months, some people still resist the changes. Humans are creatures of habit. It is hard to move them away from their comfort zones.

This type of behavior is easily observed when companies consider AI solutions. Their teams are demanding the same reports, the same forecasting methods, the same process flows. They claim to be looking for AI improvements, but their requests are basically the automation of sometimes outdated business practices.

The Bad and the Ugly

The most challenging AI projects I have implemented were the ones where companies lacked two critical things:

• A vision

• A sponsor

The need for a clear understanding of AI adoption

A clear vision of what your company wants to achieve through AI adoption will help to differentiate normal software solutions versus AI solutions. This is not a philosophical exercise. It is a highly pragmatic and efficient way to start the process. Even an initial draft of a vision allows companies to connect with experts and get their feedback BEFORE initiating labor-intensive RFPs, endless intro calls, and provider demos, only to end up frustrated.

Allocating time to develop a clear vision before procuring software solutions will yield an excellent Return on Investment. Remember:

True AI is NOT just another software implementation.

Here is a tip for getting started on building a vision for your organization’s AI adoption: Initiate by understanding the difference between traditional software and AI-based solutions.

Traditional software solutions are very good at organizing data, such as accounting systems and ERPs. However, they do not act as internal agents capable of driving business results autonomously. Meanwhile, AI-based solutions can recommend and continuously refine business decisions for you.

This concept reinforces the core message of our discussion: adopting AI-based platforms will require substantial changes in how processes and people interact.

Finally, without a vision, your team will not know where they are heading. This lack of clarity immediately creates resistance to adoption and lots of anxiety, whether openly visible or subconsciously hidden.

A vision without sponsorship is a presentation. A vision with sponsorship becomes execution.

Everybody with a few years of experience in the corporate world recognizes the importance of having an internal sponsor for transformation initiatives. An efficient sponsor ensures the cultural change required for AI transformation by acting on people, processes, and the technology itself. Above all, it ensures the execution of the company vision.

The sponsor cannot be a mid-management resource or a single IT team member. It must be a decision-maker capable of systematically driving and tracking the vision's execution across the organization.

Without an executive sponsor, there is a real risk that your AI initiative will be treated as just another software implementation, which will drastically limit its outcomes. I have seen perfectly viable projects get canceled because the company never achieved the desired ROI.

A great idea, poorly executed, is often mistaken for a bad idea.

Naturally, there are many other reasons AI projects fail. But vision and sponsorship would be my first two priorities to fix, especially considering the human factor, which is the core topic of this article.

The Good

After talking about so many potential problems, let's close on a positive note.

Throughout the transformations I experienced while leading so many AI projects, there was always something remarkable at the end of the journey.

After AI adoption, I saw talented people who had spent most of their working days stuck in boring, repetitive, labor-intensive tasks suddenly gain the most precious asset a human being can have: time.

Not only that, but AI also provided them with a continuous stream of tactical and strategic information, updated in real time.

Instead of spending their days performing repetitive tasks, they moved toward discussing tactics, challenging assumptions, and changing company strategies, resulting in millions of dollars in gains for corporations and exciting professional lives.

I helped teams move away from trying to make sense of monster spreadsheets that controlled thousands of items across their supply chains.

By acknowledging the cultural change involved, we reduced the constant friction between operations, sales & marketing, and finance teams after adopting AI-based S&OP tools.

I saw companies unlock millions of dollars in working capital through optimization, increase sales by preventing stockouts, and achieve many other operational gains.

Instead of being replaced by AI, I saw people unfold their potential and discover new career opportunities because of AI.

It was simply beautiful to watch and exciting to be part of.

Closing

We, humans, are still the best-engineered intelligence on the planet.

Yes, AI is moving fast and may eventually surpass us. Until that happens, and hopefully even after, companies that successfully combine AI tools with the massive processing power of motivated humans will lead the way and become the survivors and leaders of tomorrow.

These new market leaders are companies that fully embrace the new world of AI platforms and often achieve measurable financial and operational gains within weeks, not months or years. They are successful by merging technology with the human capital already available within their organizations, guided by a clear, transparent vision and an empowered sponsor.

Coming soon, the next chapter: The Forecasting Myth

For decades, companies built entire planning structures around forecasting. Spreadsheets became bigger, statistical models became more sophisticated, and teams spent endless hours trying to predict an increasingly unpredictable world.

But what happens when AI changes the game from static forecasting to continuous recalibration in real time?

In the next article, I will explain why many traditional forecasting methods are quickly becoming obsolete and why companies that continue to operate the old way may soon find themselves reacting too slowly to compete.