00:00:07 Introduction to Bionic supply chain concept and Stefan Gstettner’s background.
00:01:58 Explanation of the Bionic supply chain and its importance in the industry.
00:03:39 Lokad’s approach to the Bionic supply chain and the role of supply chain scientists.
00:05:36 Challenges in human-machine collaboration and the need for a Bionic operating model.
00:07:50 The four elements of a Bionic operating model: guidance, operating model, skills, and technology platform.
00:08:29 Pitfalls of transforming supply chains in the industry.
00:10:47 Executives focusing on increasing value and preparing for the future.
00:12:41 Dealing with complexities in supply chains and embracing technology.
00:15:50 The concept of a digital twin and its role in a bionic supply chain.
00:17:18 Utilizing tools to augment human intelligence in supply chains.
00:19:31 The challenge of accessing talent and the time it takes to train them.
00:20:22 The role of human interaction in future supply chains.
00:21:28 The importance of changing context to drive behavior change in supply chains.


The Bionic supply chain, a concept focused on human-machine collaboration, is crucial for the future of supply chain management. Stefan Gstettner and Joannes Vermorel discuss its importance and challenges, highlighting the need for end-to-end perspectives and human-centric operating models. Gstettner proposes a bionic operating model with four elements: guidance, operating model, skills, and technological platform. Vermorel emphasizes “approximately true” forecasting methods and better tools to augment human intelligence. Both experts acknowledge difficulties in transforming supply chains, but agree that balancing human intuition and technology is essential for adapting to the changing landscape of the industry.

Extended Summary

They discuss the concept of the Bionic supply chain and its implications for the future of supply chain management.

Stefan Gstettner provides an overview of his background, stating that he has 25 years of experience in supply chain management. He has held various roles in consultancy and has also run the operations of an omni-channel retail company in Germany for seven years. Additionally, he teaches supply chain management at MIT to help address the talent bottleneck in the industry.

The Bionic supply chain, according to Gstettner, is not meant to be another buzzword but rather a serious concept. It emerged from the BCG think tank’s work on the company of the future. The Bionic company is centered around the collaboration between machines, artificial intelligence, and humans, focusing on how to best utilize this combination. This concept is particularly relevant for supply chain management, as it is heavily reliant on technology, analytics, and artificial intelligence, as well as collaboration between humans and machines.

Joannes Vermorel shares his thoughts on the Bionic supply chain, agreeing that it is a very relevant concept. He explains that Lokad, a software company specializing in supply chain optimization, has been working on the idea of a Bionic supply chain, albeit accidentally. They initially focused on developing software tools for predictive optimization of supply chains but quickly realized the complexity of the problems they faced. The idea of having a machine learning algorithm that could autonomously solve problems like store location proved too difficult due to the vast number of use cases.

The discussion highlights the growing importance of the Bionic supply chain concept in supply chain management. The idea revolves around effectively leveraging the collaboration between humans, machines, and artificial intelligence to optimize and improve supply chain operations. Both Gstettner and Vermorel emphasize the importance of this concept for the future of supply chain management, while acknowledging the complexities and challenges that come with implementing it.

They explore the challenges of human-machine interaction and the importance of a bionic operating model for supply chain management.

Vermorel shares his company’s approach, which focuses on the role of a “supply chain scientist.” This individual is responsible for generating decisions and insights while also being the first point of contact for people who challenge the system. He notes that this role is not yet considered “bionic” but is moving in that direction.

Gstettner emphasizes that while there has been significant investment in digital supply chain management, the full value has not yet been realized. He believes that the focus on individual sub-functions has led to a loss of the end-to-end perspective, which is crucial for effective supply chain management. To address this, Gstettner argues for a bionic operating model that consists of four elements: guidance, operating model, skills, and technological platform.

Both Gstettner and Vermorel acknowledge the difficulties in transforming supply chains. Vermorel highlights the issue of problem displacement, where improvements in one area might cause issues in another. He also emphasizes the challenge of managing complexity in end-to-end software solutions. In contrast, Gstettner points out that executives are primarily interested in increasing value for their companies and preparing for the future. They are focused on organizational changes and foundational IT support to drive change and make their companies future-ready.

Vermorel discusses the importance of being “approximately true” in supply chain management, as opposed to being “exactly wrong.” He notes that humans are good at this, but machines often struggle. Lokad has made progress in this area through probabilistic forecasting, which provides a more accurate representation of the future than classical forecasting methods. However, adopting this approach has been a challenge for those working in the supply chain industry due to its differences from traditional methods.

In the present day, Gstettner observes that supply chain executives are focused on looking ahead to prepare for the future. They are interested in understanding the potential technological and organizational changes that will shape their industries in the coming years. Executives are grappling with the numerous technology opportunities available and trying to identify the best ways to drive change in their organizations.

This interview highlights the need for a bionic operating model in supply chain management, which emphasizes end-to-end perspectives and human-machine collaboration. Both Vermorel and Gstettner discuss the challenges of transforming supply chains and the importance of finding a balance between human intuition and technological advancements. Supply chain executives must look to the future and consider how their organizations can adapt to the changing landscape of their industries.

Vermorel expresses his enthusiasm for the concept of a bionic supply chain, but also acknowledges the frustrations and complexities involved in dealing with supply chain realities. He explains that, due to the messiness of the real world, supply chain management involves dealing with a vast array of statistical models and complications, such as flooded warehouses and expiration dates for fresh food. Additionally, Vermorel notes that when people face KPIs, it can be challenging for them to accept and adapt to changes that impact their personal interests.

Chandler introduces the buzzword “digital twin,” and Gstettner admits that it is a difficult concept to define. However, he suggests that a digital twin represents an end-to-end supply chain, even if it is not 100% accurate. By achieving an 80% accurate representation, supply chain managers can better understand system dynamics and make smarter decisions based on scenario analysis.

Vermorel argues that to make supply chains more bionic, better tools are needed to augment human intelligence. He cites Excel as an example of a tool that has allowed humans to perform quantitative analyses at an inhuman scale, but acknowledges that it is not a sufficient solution for modeling complex systems. Vermorel believes that programming approaches have been the most successful in turning human insights into something more automated, though he recognizes that future technological advances could prove him wrong.

Gstettner agrees that increased automation and technology will play a larger role in supply chain decision-making, but insists that human interaction will always be necessary. He emphasizes the importance of designing a human-centric operating model for the supply chain of the future, which would require changing the context in which humans behave. Gstettner believes that the key to changing human behavior is altering the context, such as creating balanced, well-synchronized sets of targets to address conflicting goals like inventory management and product availability.

Full Transcript

Kieran Chandler: Hey, the convergence of a range of evolving technologies, such as artificial intelligence, blockchain, and the Internet of Things, has led to the evolution of a new concept known as the Bionic supply chain. Today on Lokad TV, we’re delighted to be joined by Stefan Gstettner, a partner at BCG, who’s going to tell us a little bit more about this concept and how it compares to some of the existing techniques on the marketplace. So Stefan, thanks very much for joining us today.

Stefan Gstettner: Thanks for having me.

Kieran Chandler: Perhaps as a first sort of question, you could tell us a little bit more about your background and also what you’re doing at BCG.

Stefan Gstettner: Absolutely. As you said, I’m a partner in BCG focusing on end-to-end supply chain management and the associated logistics questions. In this role, I’m helping our global clients in core industries to take advantage of the new opportunities in digital technologies and gradually transform their supply chains into digital supply chains. To be able to do this, I have a 25-year background in supply chain management. I started with a PhD at a time when it wasn’t even called supply chain management. I’ve had various roles in consultancy, but I also ran the operations of an omni-channel retail company back in Germany for seven years. Additionally, I’m an adjunct professor at MIT, teaching young talents in supply chain management in their global scale program, which I very much enjoy.

Kieran Chandler: Brilliant, and today our topic is this idea of a Bionic supply chain. It certainly sounds very futuristic, but perhaps you could give us a little bit of an initial overview.

Stefan Gstettner: We’re not trying to introduce another buzzword into the game because there are already many buzzwords around. My colleagues from the BCG think tank, the Bruce Henderson Institute, have developed a view on how the company of the future will look like. It’s a lot about the collaboration between machines, artificial intelligence, and humans and how the company should organize in order to make the best use of this combination. They called it the Bionic company. Now for supply chain management, it’s particularly relevant because supply chain is about technology, analytics, artificial intelligence, but maybe even more important, it’s a lot about collaboration between humans and machines. So we call it Bionic supply chain, without trying to introduce a new buzzword, but as a reasonable theme and a good headline for what we want to achieve.

Kieran Chandler: We absolutely love a buzzword. Joannes, we’ve spoken in the past about the relationship between humans and machines when we looked at user interfaces. What are your initial thoughts on this idea of a Bionic supply chain?

Joannes Vermorel: I believe it’s very relevant, and it actually matches quite a lot with what Lokad has been doing, although a bit accidentally. We didn’t have this big master plan or vision. The way we approached solving supply chain problems is by developing software tools for predictive optimization of supply chains. We quickly realized that there was so much complexity that we had to cope with. The idea of having a machine learning algorithm that would figure out on its own how to solve a problem, as simple as store allocation, was just too much because there are too many edge cases. So we needed to have something…

Kieran Chandler: So Joannes, can you tell us more about Lokad and what you do?

Joannes Vermorel: We started with the idea of supply chain optimization where we could start from human insight and shape the problem into something that the machine could use to scale out the insights at the scale of a large retail network. That was the idea behind what we did, and it ultimately became this sort of idea of what we call at Lokad the “supply chain scientist,” which is literally someone who is frontally responsible for crunching the data and generating decisions and insights while being the first point of contact for actual people who can challenge whatever is happening in this system. So, yes, I think we are maybe not yet to be qualified as Bionic, but it’s a path in this direction. Although I didn’t have something as sharp as “binding supply chain” to guide our path when we were working for the world.

Kieran Chandler: Historically, why would you say it’s been so difficult for humans to interact with machines?

Stefan Gstettner: I wouldn’t even say it’s historically. We probably all have already coded algorithms 20 years ago also, so that’s not new. But, in the past years, there have been a lot of investments into digital, let’s take demand forecasting, for example. There’s a huge amount of intelligence flowing into demand forecasting, and then there is obviously also collaboration between humans and algorithms here. However, the executives I talked to don’t see yet that the full value out of the supply chain has come. For my insight, it’s because the humans were quite busy in trying to understand how can I collaborate with a machine? It’s much more demanding to do it with algorithms than to do it with Excel, for example. Now being so busy doing that, I think what has been lost is that end-to-end supply chain management is about end-to-end. Surprisingly, it’s not about focusing on a sub-function. It’s about aligning and synchronizing the end-to-end supply chain, and that, in my eyes, has lost a little bit of focus. So, now we have to enter into the Bionic type of thing to re-emphasize that humans need to collaborate, obviously with the help of machines, and that is what has not happened yet. And that is why some executives are frustrated because the investments haven’t paid off. And then, therefore, we are saying it needs a Bionic operating model which is essentially four elements: first of all, the guidance in the supply chain, so the good old KPIs, what do we want to achieve with our supply chain? Is it speed, is it reliability, is it cost focus? And how do we segment the supply chain, and how do we put target values to it? Good old supply chain management, nothing digital without it or in it. Then, the operating model, which is in our belief a very much platform-driven operating model to put the people who are synchronizing the supply chain together, even physically together sometimes. Secondly, skills, of course, I mean, a totally different set of people who are capable of designing the entire and supply chain. And then, the technological platform, but maybe not as the first one, but as the foundation of course, we need it. And that in very few companies has been established so far. And therefore, I think historically, that is a bit of the disappointment about the digital supply chain.

Kieran Chandler: So, Joannes, as a software vendor, what do you think are the pitfalls that companies face when trying to optimize their supply chain?

Joannes Vermorel: Well, as a vendor, if you don’t manage to kind of grab this end-to-end, what you actually effectively deliver is just problem displacement. You displace a problem from one location to another, and maybe your KPIs on the silo that you’re taking over look good, but actually, you’ve created problems sideways. So, you’ve not actually delivered much worse skills. You’ve actually destroyed values, which is very, very bad. Unfortunately, some large enterprises are quite good at doing that. And then the challenge is that as soon as you want to do something like end-to-end software-wise, it becomes such an incredible challenge because your software becomes so incredibly complicated. And I think most people would agree that the average ERP is, we are talking thousands of tables, hundreds of screens. So, it’s like, whoa, so much complexity, and we are barely scratching the surface. So, we have really a problem of complexity management. The question is, how can you get what people are doing in Excel? They are doing something that is very human, but that’s very, very difficult to replicate with machines. It’s being approximately true as opposed to being exactly wrong. Machines are very, very good at being exactly wrong at a large scale. So, at Lokad, we had some mini breakthroughs, like probabilistic forecasting, to get closer to this idea of approximately true, you know, state of operation. But it’s an ongoing change. And the thing is that while I think it’s a technologically very relevant answer, it’s also a very surprising answer to people who are running the supply chain because it doesn’t look at all what they were doing before, especially compared to classical forecasting. So, probabilistic forecasting is forecasting. It’s about knowing things about the future, but it’s so bizarre compared to the classical way that indeed in terms of adoption, it’s a big chain on its own.

Kieran Chandler: Okay, Stefan, how are supply chain executives approaching these problems and trying to implement new technologies?

Stefan Gstettner: Well, executives are not primarily interested in implementing technologies. So, they are interested in increasing the value for the company and preparing for the future. What they are seeing increasingly and what they get external input on all the time is that, of course, in ten years from now, it will all look massively different. So, my favorite saying, I think it was Gates who said this, we are always overestimating the change in the next two years and underestimating the change in the next ten years. And the role of executives is to look ten years ahead and not only two years. And this is what they are now asking their organizations to do. The future will look totally different. Let’s face it. We don’t see it now, but it will be the case. So, how can we change, not primarily technology-wise, but then also organizational-wise? That’s why we are emphasizing the operating model more. And then, of course, they are also asking, what will be the foundational platform or the foundational support on the system and data side and analytics side in order to drive this change? And they are still struggling to understand why does my organization not drive the change that I want it to change? There are some inhibitors who still don’t make the company future-ready. So, that’s what’s in the top of the mind of the executives.

Kieran Chandler: Let’s talk about a bionic supply chain. It sounds like something out of Ironman - very cool and futuristic. What does it actually look like in reality?

Joannes Vermorel: The reality of a bionic supply chain is both frustrating and fascinating. The frustrating part is when you want to cope with the supply chain reality, you end up dealing with hundreds of tables that represent a digital counterpart of the world. The world is very messy, and if you want to be accurate, you need to cope with this complexity. You end up having to deal with statistical models to address things like flooded warehouses, bizarre compatibility issues in aerospace parts, or expiration dates at the lot level for fresh food. So, there are many complexities that need to be embraced, and it takes a lot of effort. Tech-wise, you don’t end up with something as pure or clean as most good software products can be.

There’s another aspect too. In theory, we’d like to have a super rational approach to supply chain optimization, where people would agree that if it improves the bottom line, it’s good for the company. If there are externalities like environmental impacts, we would factor that in so that we properly optimize, including externalities. But in a very mundane way, when people face KPIs, it’s tough on them.

For example, we had an experience with a large retail network where we diagnosed that the stock in the stores played two roles. One is to serve clients, so you have some inventory to ensure that when people walk into the store, they can be properly serviced. You then realize that a significant portion of the stock is not there for this purpose, but for merchandising purposes – to make the store appealing. You might conclude that you should move the budget for the stock meant for merchandising purposes to marketing instead of supply chain. However, this can lead to intense political struggles within departments that are suddenly faced with a massive amount of stock being pushed onto their budget. This can sometimes affect bonuses and create resistance to change.

Kieran Chandler: As we love buzzwords, let’s introduce one more: the concept of a digital twin. How does that fit in with the idea of a bionic supply chain, and where does it come into the picture?

Stefan Gstettner: In my experience, the term “digital twin” can be difficult because everyone has a different understanding. It’s already challenging to understand supply chain, but adding the word “digital” makes it even fuzzier. When you have a digital twin, it seems like nobody knows what it’s about. I won’t attempt to define a digital twin, but the idea is to have a representation of the entire supply chain. I fully agree that it will never be 100% accurate, but right now, we have only a 10% accurate end-to-end representation. If we can achieve something like an 80% accurate representation of the end-to-end supply chain, we can start learning about the system dynamics. We could understand the implications of a spike in demand forecast, how to adjust target inventory, how to play around with production sequences, and what it means for the supply side. We can also explore different scenarios. In this sense, a digital twin can be the vehicle to enable people to think in this way.

Kieran Chandler: Who are running the end-to-end supply chain to play around with scenarios and take smart decisions, and if that is meant with a day-to-edit wind, do you buy into this as one of the key enablers?

Stefan Gstettner: Yes, I do believe that this is one of the key enablers for smarter decision-making in supply chains.

Kieran Chandler: Joannes, if we look ahead to the future and what we do to our supply chains to make them a little bit more bionic, do we need better classes of tools to augment human intelligence?

Joannes Vermorel: Yes, augmenting human intelligence is crucial for supply chains. Even though it’s a big word, it can be something simple. For example, Excel allows a person with average mathematical capabilities to deal with a large amount of numbers. In the nineteenth century, it would have been impossible to perform thousands of basic operations per day with just a clerk, even if they were brilliant. So, in a way, tools like Excel give people quantitative insights at a relatively inhuman scale.

However, Excel has its limitations when it comes to modeling complex systems, so it cannot be the end answer. In my opinion, the most successful approaches so far have been programming-based, as they help decouple human insights and turn them into something more automated. I think this trend will continue for at least the next couple of years, but it’s possible that I’m underestimating some paradigm shift that might happen later.

I also know that some people at MIT are working hard to deliver smart talents to the supply chain field. However, I believe that getting access to all the talent needed will be a significant bottleneck for multiple decades. It takes a long time to train people who can, in turn, train others. That’s one of the reasons why Silicon Valley remains so successful; they’ve had generations of talented software engineers who can train more software engineers and produce more innovative companies. The same goes for MIT – many universities are trying to replicate its success, but it will take decades to get there.

Kieran Chandler: Stefan, in your view, what would a successful Bionic supply chain initiative look like?

Stefan Gstettner: I agree that there will be a much higher degree of automation and technology will take over a large percentage of decision-making in the supply chain. However, I think there will always be a need for human interaction. We need to design a human-centric operating model for the supply chains of the future, which includes the elements I described earlier.

One aspect to consider is human behavior. If we want to change behavior, we need to understand how to do so, as the historical silo behavior hasn’t been particularly successful. We believe that the context in which humans behave is the most important driving force for behavior. If we don’t change the context, for example, by setting balanced and synchronized targets, then humans will never change their behavior, and we will never drive success and performance.

I think the smart combination of what companies like Lokad are doing and what companies need to acquire on the human side is the key. This connection between the two, which we might call the Bionic supply chain, is the future.

Kieran Chandler: Brilliant. We’re going to wrap it up there. Thanks for your time, both of you.

Joannes Vermorel: Thank you.

Stefan Gstettner: Thank you.

Kieran Chandler: Thanks for tuning in this week, and we’ll see you again next time. Bye for now.