Three ways GenAI could transform supply chain management
GenAI is becoming a more common feature of the supply chain function, but significant areas of opportunity remain, suggests IMD’s Carlos Cordon. ...
Artificial Intelligence
Two data points and four keys to keep in mind with AI
Most AI projects are destined to fail. Here are four tips to avoid falling into that trap.
The artificial intelligence (AI) earthquake has taken many business leaders by surprise. Impressed by generative AI systems, such as ChatGPT, executives are seeking this technology, hoping to generate and capture business value. By the end of 2026, 80% of companies are expected to be using, or will have developed, some kind of GenAI product or service.
While 80% is a high figure, it’s not surprising if we consider what AI-based systems have achieved in the last two years alone. They have solved mathematical problems that had gone unsolved for decades, aced bar exams, and diagnosed diseases with error rates lower than human experts.
However, implementing AI in a company successfully is not easy. In fact, it is estimated that 80% of AI projects are destined to fail. The dilemma facing business leaders today is clear: How might AI be leveraged to generate and capture value without being destined for the bin, like the majority?
My answer to this question has four parts. First: thoroughly understand what this technology is and how it works. Second: clearly identify the specific problem to solve using AI and, thereby, the value that AI can bring. Third: determine precisely what data is needed to feed the AI so that it can effectively address the identified problem. Fourth: anticipate and manage the reactions of people within your organization. My advice to executives is broken down into the following four steps:
To drill down to the basics, AI is based on machine learning models and large language models (LLMs)
1. Understand that AI is a tool for pattern recognition
To drill down to the basics, AI is based on machine learning models and large language models (LLMs). Machine learning models analyze data to solve specific problems, such as distinguishing between images of healthy cells and cancer cells to aid in medical diagnoses. What lies behind that learning process? In essence, it’s sophisticated statistics. LLMs, such as those used by OpenAI’s ChatGPT, take large amounts of text to generate new content, predicting which words have a high probability of forming sentences that give coherent responses to prompts entered by users. This, too, is sophisticated statistics. Both types of AI analyze data to learn patterns and deliver results that are as likely as possible to meet users’ needs. Nothing more, nothing less.
In marketing, AI can personalize advertising campaigns by analyzing customer behavior.
2. Identify problems whose solutions generate real value
Identifying real problems that can be solved with AI to generate value for a company is a necessary condition, though not a sufficient one, for successful implementations. Don’t fall into the trap of looking for AI “use cases” (a mantra repeated all too often in companies today). Take time to understand the specific needs that can be served effectively. For example, a bank can use AI to detect fraud by analyzing unusual patterns in transactions. In marketing, AI can personalize advertising campaigns by analyzing customer behavior. In a factory, AI can optimize the supply chain by predicting machine failures and managing proactive maintenance. Identifying real problems to solve is the only way to implement AI so it generates tangible value.
3. Recognize the importance of data
Training AI with proprietary data ensures that the system is tailored to your business’s particular needs, thereby increasing the odds of success and generating greater value for your business. AI models, as we mentioned earlier, recognize patterns in data; company-specific data is therefore essential for the model to learn in an accurate and relevant way. In the case of the bank, historical transaction data will help the AI identify fraudulent behavior. In the marketing example above, customer interaction data will help personalize campaigns. In the factory example, machine maintenance and performance data will facilitate failure prediction and help make decisions that optimize production. Even if it may seem like a cliché, it is good to remember GIGO: garbage in, garbage out. Meanwhile, rich and good quality data is the way to rich and good quality AI results.
4. Focus on people
Finally, it is critical to remember that, even in the context of business, AI is a tool to help people. While AI can analyze large volumes of data and find hidden patterns, people bring context, creativity, and critical judgment. In a bank, human analysts interpret and act on AI-generated fraud alerts. In marketing, creative people design strategies based on AI recommendations. In a factory, human engineers use AI predictions to improve processes and make decisions. The synergy between AI and people maximizes the technology’s potential and ensures its acceptance and success in the organizational environment.
It is only when human intelligence uses AI to expand our capabilities to achieve useful results that we can consider AI well deployed.
Intelligent machines?
Alan Turing, considered one of the founding fathers of AI and modern cognitive science, envisioned a future where machines could think and learn. However, it’s important to keep in mind that so-called artificial intelligence is nothing more (or less) than a family of sophisticated data processing tools. While it’s true that, in some contexts, these tools may appear to be intelligent, AI in and of itself will not generate business value for us. It is only when human intelligence uses AI to expand our capabilities to achieve useful results that we can consider AI well deployed. It was Turing himself who said: “We can only see a short distance ahead, but we can see plenty there that needs to be done.” That maxim is still valid today and should invite us to continue to deepen our understanding of this technology while experimenting to find solutions that generate real value.
AI x 9: This article appears in a nine-part summer series that examines how AI is impacting leadership and business, produced in collaboration with Expansión.
Authors
Related
Recruitment in 2025: AI is a great aid, but don’t forget the personal touch
AI is rewriting the rules of hiring – job seekers and recruiters alike must adapt to a process driven by algorithms but grounded in human connections. ...
IMD’s AI Safety Clock ticks closer to midnight
Breakthroughs in agentic AI, open-source development, and closer collaboration between AI companies and the defense sector have significantly raised the AI risk profile. These developments have compelled us to advance the AI...
Focus on data governance to reap the rewards of GenAI
GenAI can deliver substantial benefits, but few organizations understand the risks. They need a deliberate data governance strategy....
Learn Brain Circuits
Join us for daily exercises focusing on issues from team building to developing an actionable sustainability plan to personal development. Go on - they only take five minutes.
Explore Leadership
What makes a great leader? Do you need charisma? How do you inspire your team? Our experts offer actionable insights through first-person narratives, behind-the-scenes interviews and The Help Desk.
Join Membership
Log in here to join in the conversation with the I by IMD community. Your subscription grants you access to the quarterly magazine plus daily articles, videos, podcasts and learning exercises.