Pakorn Khantiyaporn/123RF

What Challenges Face AI Adoption In Supply Chain Management? 19 Experts Share Their Insights

  • 27 September 2019
  • Sam Mire

Supply chains breed headaches. From product loss to unreliable vendors and suppliers, supply chain managers have enough on their plates. And while AI promises to simplify the overall supply chain landscape, it will present its own set of challenges. This industry insiders shared their thoughts on what those challenges are and how they'll impact AI's adoption by supply chain managers. Here's what they said:

1. Trevor Stansbury, founder and CEO of Supply Dynamics

“The #1 challenge in relation to most manufacturing supply chains is the absence of data governance policies that are enforced and adhered to across an OEM’s extended supply chain.

What most people don’t understand is that over the past 30 years, the vast majority of manufacturing companies have outsourced most of what they once made in-house.  As crazy as it sounds, General Electric doesn’t make engines.  Caterpillar doesn’t make heavy earth moving equipment.  Boeing doesn’t make airplanes. And Ford doesn’t make cars.   Their Tier 1-Tier N suppliers do. Sure, the OEM designs and assembles the finished product but “the chips” are often cut somewhere else.

Unraveling all of that can be maddening and since AI depends on large data sets that conform to a common naming convention and data taxonomy, adoption of AI has been slower than one would predict.   In order to exploit machine learning, NLP, blockchain or many other technologies in the manufacturing supply chain, OEMs and their tier 1-Tier n supplier will first have to agree on what something is called, how and where it is recorded and securely shared across an enterprise.   In that sense, how you pronounce tomato (“tomayto” OR “tomahto”) really does make a difference.”

2. Richard Lebovitz, President and CEO of LeanDNA

“The top challenge for AI adoption is building trust in the output and recommendations. There are natural hesitancies to believe machines that don’t have years of experience in manufacturing and supply chain. It’s tricky to combine data and new algorithms with practiced human domain expertise—people need to see how the machine came to its conclusion to make sure they can trust the results.

That’s why those incremental steps to AI development are pivotal. Once data is visible and accessible to teams, technology must demonstrate to them that the automated recommendations are reliable and backed by best practices. Without that trust and context, teams simply won’t adopt new technology. They will continue to do it the way it’s always been done—manually through time-consuming spreadsheet reporting.”

3. Rajeev Gollarahalli, CBO of 42Q

“Lack of high-quality and consistent data compromises AI solutions but as you feed more and more data into a machine learning platform, it can make better decisions. Once you solve the data puzzle, your organization must also make the cultural shift to AI. For decades, supply chains have been managed primarily based on tribal knowledge. Moving from a mindset of ‘this is how we’ve always done it’ to ‘this is what the data is telling me,’ is a newer concept that is essential for the successful adoption of AI in supply chains.”

4. Dan Patt, CEO of Vecna Robotics

“Accessibility, and scalability. People tout flexibility as the key to supply chain success. Yet, the process to buy, integrate, and scale these solutions is long, intricated, and costly.

Over the next few years, companies will move away from one-time purchasing models, even away from Robotics as a Service, and towards the tried and true SaaS model – taking the hardware out the equation completely to focus on the software’s core abilities. By moving from a steep up-front cost to a nominal monthly charge, companies can approve AI initiatives faster and recoup ROI quicker. This allows them not just to buy a piece of equipment today, but invest in the future with continuously improving capability.”

5. Govin Ranganathan, Senior Manager of Logistics at NIO

“AI solutions also require significant investment and monitoring, and thus there is senior management foresight and patience required, and eventually benefits will outweigh investment. AI solutions are more likely to be black box than traditional solutions. With black box solution, material planners are unable to understand how the process, method, and algorithms work, which creates issues with the accountability of results.”

6. Anand Medepalli, Head of Product at Element AI

The biggest challenge to AI adoption in supply chains is data silos. For example, it is quite common for the marketing department to have their own database of customer information, as does a store manager. AI provides the ability to sift through massive amounts of data and convert that into relevant information, but in order to do that, data siloes must be broken down, communication improved within the organization. A businesses’ ability to harness digital threads and distinguish what it is the customer is looking for – through this process of sifting through data – is part of that challenge.”

7. Paul Noble, CEO of Verusen

“The No. 1 challenge to AI adoption in supply chains today is like most transformations―ensuring the right stakeholders are aligned on the objective. Too often, there's a misconception of what AI can do, and though, there's significant potential in AI, it still requires humans to do the upfront work and set it up.

In many ways, AI can be like a human where initial deployment means it's like an infant. It requires transformed/formatted data. It requires training. It requires time to improve and continually optimize for the task at hand. Of course, if there's an ability to learn quickly through a network, some of this can be mitigated with strong results.”

8. William Crane, founder and CEO of IndustryStar

“Directly put, there are many companies that have an amazing piece of developed AI software technology whose teams are searching for a supply chain problem to solve. AI, as with any technology, is a tool to do a job. The secret to AI success is understanding what tool is needed to solve a specific business problem.”

9. Heather Gadonniex, the VP of Marketing at Samasource

Forrester estimates the AI market will reach $1.2 trillion by next year, but the technology itself has limitations that need refining. McKinsey & Co. lists the top five limitations as, labeling training data, obtaining data sets, explainability, carrying learnings between models and bias.

Companies face challenges obtaining enough AI training data, developing strategies for robust data quality and ensuring that bias does not occur. Additionally, companies must ensure the labor force and technology they are using for labeling are ethically sourced.”

10. Ed Clarke, co-founder and Managing Director of Yojee

“Everyone is still learning, which is why we take a very involved approach to selling and implementing our product.

We have our business development managers set up bespoke demos for potential clients so that they can see how their particular use case can be managed through Yojee.

We also have on-site training for dispatchers and drivers, plus a dedicated home office team that manages our in-product support channel, and for our larger clients, we have seasoned project managers working alongside them to ensure smooth and efficient roll-outs.”

11. Oren Zaslansky, CEO of Flock Freight

“I think it's a combination of trying to apply new technology to legacy supply chains (typically diminishes efficacy) as well as applying AI in totally novel and yet unproven models. It's still early days and successful applications are still shaking out.”



12. Dr. Madhav Durbha, Group Vice President, Industry Strategy at LLamasoft Supply Chain Management Software

“Humans are the #1 challenge. Part of the resistance comes from the insecurity of giving up control and trusting algorithmic decisions. Supply chain professionals tend to think in logical causations than trusting correlations to drive decisions. AI can be seen as a black box in this regard. However, as explainability and detection of bias in AI advances in other industries such as finance and banking, SCM will benefit. Another part of the human resistance comes from a Luddite fear of algorithms taking over human jobs. Those who embrace the new reality of augmented decisioning powered by AI will emerge as winners.”

13. Pervinder Johar, CEO of Blume Global

“Perhaps unsurprisingly, the biggest challenge of AI adoption is almost always the human element and ensuring the team is educated on the necessary skills and tools, which can sometimes require retraining employees. While the application of AI enables a way to enhance efficiency, understand data, drive decision-making and accelerate actions, it’s not meant to be a substitute for good relationship management by the supply chain team. Bottom line: AI is not a replacement for human judgment. People are vital to the successful implementation of any automated system — and in supply chains especially, AI should never work in isolation.”

14. Rajesh Kalidindi, founder and CEO of LevaData

The obvious answer is data and the lack of the four “Vs” (volume, veracity, variability, and velocity). In global supply chains, this is complicated by the fact that so much of the data lives outside the enterprise, shared or “owned” with suppliers, channel partners, logistics service providers, and third-party manufacturing partners. 

Because solutions to address data issues are being widely adopted, however, the bigger challenge may be human factors – adjusting to change and developing trust in the predictions and recommendations that AI can generate.”

15. Jake Rheude, VP of Marketing at Red Stag Fulfillment

“I think we see a lot of companies — this goes not just for supply chain, but for other industries too — tout how they're using next-gen AI to streamline their business. And sure, sometimes there's real progress that comes from adoption of AI instead of relying on manual calculations or other human labor.

However, there are plenty of cases where AI gets used not because it actually is going to make a difference, but because it looks good in shareholder reports, or gets free PR (a big example being blockchain + anything else back in 2017). I think the biggest challenge is that AI adoption should happen because it will make a positive difference to a company's operations, and not just because it makes a splash.”

16. David Hogg, Vice President of Business Development at Logistyx Technologies

“Adopting AI as part of the supply chain can open the door for potential legal complexities, data and privacy breaches, and identity theft. The more data companies can harness through AI, the greater the possibility their supply chains become targets for security breaches, especially in multimodal shipping with many players involved. That's why it's essential to have a multi-carrier shipping system with advanced encryption, access control, and continuous penetration testing to ensure data remains safe.”

17. Kamal Anand, Chief Technology Officer of Bamboo Rose

“Many people use the excuse that there is “not enough data” for AI adoption in the supply chain. However, this is untrue – companies have the data. The main challenge is companies have been running the same way for decades and data is living in different places and systems – from emails, to reports, to orders, and even on paper. The question becomes, how do you unify your data?

This makes true collaboration a challenge. In order to make changes to the supply chain on the fly, companies must adopt digitization and keep their community of manufacturers, shippers, suppliers and more, connected, which is imperative during the current trade chaos. Without this seamless integration, AI adoption in the supply chain becomes extremely difficult.”

18. Doug Surrett, Chief Product Strategist at BluJay Solutions

“One of the biggest challenges is to determine the “what” before the “how” – i.e., rather than looking for ways to use the coolest technology, start with the end goal and then determine how technology can help accomplish it. For example, in the context of autonomous vehicles or drones, accuracy is critical to pinpoint addresses and deliver products to consumers without mistakes. Geocoding technologies have the potential to overcome this hurdle.

Another major challenge will be changing mindsets to adapt to change rather than fear it. For example, AI is not likely to eliminate the need for drivers, however their roles may shift if autonomous vehicles gain widespread adoption.”

19. Emily Murphy, Editor of Supply Chain Brief

“The biggest challenge in AI adoption is accepting change. The implementation is very beneficial to Supply Chain leaders and mangers, who are working with high volumes of data every day. But letting go of traditional tools like spreadsheets and allowing more complicated planning systems to take over is a big change and requires these managers to trust AI.”

Have expert insights to add to this article?

Share your feedback and we'll consider adding it to the piece!


About Sam Mire

Sam is a Market Research Analyst at Disruptor Daily. He's a trained journalist with experience in the field of disruptive technology. He’s versed in the impact that blockchain technology is having on industries of today, from healthcare to cannabis. He’s written extensively on the individuals and companies shaping the future of tech, working directly with many of them to advance their vision. Sam is known for writing work that brings value to industry professionals and the generally curious – as well as an occasional smile to the face.