AI has promised to transform nearly every industry, but it’s hit some obstacles along the way. Reports of canceled projects and costly implementations have some wondering: Will AI’s promise live up to its hype? In this episode, we’ll explore one area where AI implementation is accelerating: supply chain. We’ll discuss what makes AI in supply chain unique, how companies are investing in innovation, and what weak spots still linger, with insights from recent research conducted by Economist Impact.
Guests:
Bob Ferrari, Managing Director, Ferrari Consulting and Research Group
Oliver Sawbridge, Senior Manager, Trade and Geopolitics, The Economist Group
Episode transcript:
Clip: I will get bleeped out, but let me say it: We're always bitching about why can't we get this data sorted out? Well, look inside. This is why.
Sarah Harkins: This is Trade Secrets, a podcast from Kinaxis where we ask big questions about our world and find surprising answers by following the hidden networks that connect everything: supply chains. I’m your host, Sarah Harkins, and today we're tackling a question that's on everyone's mind: Where will AI make the greatest impact?
Every tech company, thought leader, and conference keynote is promising artificial intelligence will revolutionize our work. But the transformative results have been slow to arrive. In fact, some studies say that as much as 42 percent of initial AI projects have already halted.
So today, let’s cut through the noise and ask a specific question: Which industries are actually positioned to benefit from AI? And, more importantly, could supply chain be one of the biggest winners?
Act I: The reality check
SH: We’ll start with some hard truths: AI hasn't lived up to its potential in a lot of places. Consider the example of self-driving cars. For years, we've been hearing that they are just around the corner. And while the technology is in place in some cities, it’s not transforming our world. There are plenty of other examples of AI falling short of its promise in retail, healthcare, finance, and other sectors. So, why hasn't AI delivered on the hype?
Part of the reason is that it requires three things to work: high-quality data, clear metrics for success, and an environment where automation can integrate smoothly. In many industries, those elements are missing. But one that checks all those boxes? It’s supply chain.
Act II: Why supply chain is different
Bob Ferrari: Supply chain teams were the first adopters of AI/machine learning-based technology. We've done it in planning, we've done it in execution, and we've gotten very positive results because of it.
SH: This is Bob Ferrari.
BF: I’m managing director of the Ferrari Consulting and Research Group. My background, it’s been in all areas of supply chain, and I’ve seen a lot of things and learned a lot of things that I can share.
SH: We’re talking to Bob because he’s seen first-hand the strengths, weaknesses, and opportunities for supply chains when it comes to AI. He’s ready to help us separate fact from hype.
We also have data. The Economist Group recently surveyed over 800 supply chain leaders about their own AI experiences. They found something surprising: 71 percent say their companies are accelerating AI deployment in response to recent disruptions, like global trade policy.
71%
So, while other sectors are reconsidering their AI investments, supply chain is doubling down. That strategy makes sense when you consider the state of world affairs. According to the Geopolitical Risk Index, today’s turmoil echoes other periods of disruption and heightened geopolitical tension, like the end of the Cold War, The Gulf War, 9/11, and the start of the war in Ukraine.
Oliver Sawbridge: Other than those periods, we've not lived in a time with more geopolitical uncertainty in the last 40 years than today.
SH: This is Oliver Sawbridge, Senior Manager at The Economist Group, specializing in trade and geopolitics.
OS: Businesses now face not just new tariffs and sanctions, but unpredictable swings in policy from industrial strategy to export controls to new data governance rules. The combination of geopolitical shocks and policy volatility is undermining predictability, leading firms to build contingency into their operation.
SH: Supply chains that once focused on resilience and bouncing back from the occasional disruption have realized that process can no longer keep up. They have to adapt in the moment if they want to thrive. This is what makes AI capabilities essential.
“Unfortunately, a lot of groups umbrella all of AI, but in a classic sense, there is a narrow AI. There are AI assistants. There are AI agents. When you get into all those iterations, that requires added education.”
OS: AI can help organizations manage complexity and make end-to-end decisions more efficiently. The data from the research shows that more than three quarters of companies have integrated AI into different solutions such as predictive analytics, real-time decision making, and supplier monitoring with the expectation that it will have a high to significant impact on their business operations.
SH: During previous disruptions, companies were limited by the volume of data they could analyze and the speed at which they could do it. Now, with AI, pattern recognition has advanced, automation has accelerated decision making, and processes can continuously improve. This has created opportunities for optimizing multiple areas of the supply chain.
BF: On the planning side, the notion of demand management can be shifted towards demand sensing, sensing what's going on in the fulfillment channels and what that means. There are AI tools that can help in that area.
Supply planning is another one because one of the constant challenges is multi-tiered visibility relative to supply. We've come a long way in technology support in that area, but trying to spot patterns on those, again, gets into some challenging areas.
A third one is procurement. One of the most labor-intensive aspects on that side is supplier onboarding because you have to profile a lot of information relative to a new supplier. It's time consuming. There are certain promises of AI enablement that are going to help us with doing that.
SH: And those are just a few of the possibilities for AI in supply chain. That means the challenge for many organizations is deciding where to begin. Bob says that also means defining what AI signifies before making any investments.
“Non-executives see how difficult AI integration really is. They cite fragmented systems, poor data quality, and siloed decision making as the biggest barriers.”
BF: The term AI is all inclusive and, unfortunately, a lot of groups sort of umbrella all of AI, but in a classic sense, there is a narrow AI, which is machine learning technology, and that's been around about five to 10 years. There are AI assistants. There are AI agents. There is large language models.
When you get into all those iterations, that requires added education, the sense of what are the strengths and weaknesses of each, and also what are the foundational needs that we need to have as an organization to be able to leverage those?
SH: And this education is threatening AI advancement for many organizations today…
Act III: Setting realistic goals
OS: Our research shows that two thirds of C-suite leaders expect a financial return from AI within 12 months of adopting it… versus only 45% of modern junior leaders. Non-executives see how difficult AI integration really is. They cite fragmented systems, poor data quality, and siloed decision making as the biggest barriers.
67%
SH: This gap between leaders and planners shows how our greatest expectations for AI often conflict with the realities of planning and implementation.
BF: The gap is to be expected because at the operational level, supply chain teams of various areas have been dealing with these constant challenges of trying to bring information together from disparate systems. The testimonial to that, unfortunately, has been the high levels of spreadsheets. To get to all of the information needed to do all of the things that need to be done on a short timeframe, you literally have to convert the information from one system to the other.
SH: And it’s not just siloed data that create challenges. As Oliver mentioned, it’s also people and systems.
BF: Each function has their own performance metrics, which they get compensated on. And, well, I don’t know—I will get bleeped out, but let me say it—we're always bitching about why can't we get this data sorted out? Well, look inside. This is why. A foundational step to advanced AI is organizational related.
SH: If leaders have supply chain backgrounds, they’re likely to understand these challenges. But for those that don’t, their optimism might prevent them from recognizing real hurdles to implementation.
And even if they align with managers on execution, a year may not be enough to see real returns.
OS: Systems take longer than that to mature, and confidence can quickly erode. Managers closer to implementation tend to understand the learning curve required before AI delivers its sustained impact.
SH: And for more advanced AI applications, like agents, the technology itself may need to go through an onboarding phase before companies can feel confident in its results.
BF: Companies that I've talked to say it may be three to six months for one agent to come up to speed, to learn the process, to learn the workflows, to learn the decision points, to learn the trigger points, and then the next agent gets launched and then it learns the next process.
SH: The good news is that the survey shows that both the c-suite and managers have similar visions for AI.
OS: The gap between executives and non-executives does narrow over the medium term. So, both executives and managers expect AI to significantly improve forecasting, scenario planning, and geopolitical visibility within three years. So, optimism and realism eventually align, but only if organizations can bridge the gap between vision and capability.
SH: Implementation strategy will be a key part of this alignment, particularly around defining success.
BF: What we're hearing about almost monthly now is a lot of organizations have started proof of concepts of a lot of these AI technologies right now to gain an understanding of what are they, what are their capabilities, do we have applicability here? And let's see if we can do a couple of initial activities to garner the results.
“Advanced AI, we’ve got to be realistic. There's going to be a cycle. It may be more condensed than the five to 10 year. We don't know yet because it's still evolving.”
SH: As part of those trials, companies will also need to assess the risks.
OS: Fewer than one in four business leaders believe AI will introduce new risks, such as cyber vulnerabilities, compliance failures, or flawed risk assessments. So, firms may be overestimating AI's readiness and underestimating their own exposure.
SH: That finding is surprising given that a separate survey found at least 13% of companies have already reported breaches of AI models or applications.
BF: At the biz level in terms of what are the top five risks for businesses today, cybersecurity is number one and number two. It just comes up every single time and for very valid reasons because it is literally now every month in the news cycle we hear something about this.
SH: That’s not to say that AI’s risks outweigh the benefits for supply chains, but the future belongs to the informed optimists, the people who can look at AI with clear eyes, understand its promises and its limitations, and implement it strategically.
BF: It's like any other journey. It's going to take time. We often forget what happened in COVID when all the businesses literally said, oh my God, what are we going to do?
Well, everybody got innovative, and through that action, they discovered the benefit of cloud-based systems. Cloud-based systems existed before COVID, but the reality and the benefit hit home during COVID.
AI, advanced AI, we’ve got to be realistic. There's going to be a cycle. It may be more condensed than the five to 10 year. We don't know yet. Because it's still evolving.
Act IV: Conclusion
SH: So, to go back to our original question, where will AI make the biggest impact? And could it be in supply chain? The answer is in the hands of supply chain leaders and managers.
We've seen AI struggle in sectors where the data is messy, the goals are ambiguous, or the technology doesn't fit smoothly into existing systems. But supply chain has the potential to be different. The fundamentals are there: the data exists, the problems are well-defined, and the industry is ready for this kind of innovation.
Achieving AI’s highest potential will require setting realistic expectations and making ambitious adjustments to operations, but the opportunity is there to be seized. And if the survey data is any indication, the first supply chain team to demonstrate AI’s transformative potential has already started its journey.
Act V: Closing notes
SH: Thanks for listening to Trade Secrets, a podcast from Kinaxis. Special thanks to Bob Ferrari, The Ferrari Consulting and Research Group, Oliver Sawbridge, The Economist Group, Suhas Sreedhar, and Josh Caldwell.
If you’re interested in learning more about AI in supply chain, checking out the original research from Economist Impact, or learning more about the Ferrari Consulting and Research Group, you can visit the links in our show notes.
And one last thing. One of the many uses for AI in supply chain? It could reduce shipping delays—by avoiding whale collisions. Researchers have been testing an AI-powered detection system that gives crews enough time to change direction to reduce whale deaths and injuries. Sounds swell.
Articles mentioned and additional resources:
Supply chain’s big bet on AI for geopolitical resilience – Economist Impact, sponsored by Kinaxis
Homepage - The Ferrari Consulting and Research Group
Supply Chain Matters – The Ferrari Consulting and Research Group
AI project failure rates are on the rise: report – CIO Dive
Woods Hole and Matson Test Out Real-Time Whale Detection System – The Maritime Executive
