Navigating the Supply Chain AI Fog

Understatement alert: this summer has seen mixed messages and sentiment regarding AI. A fog has rolled in, and it feels like we’re all trying to make our way through it without a sense of how the broader landscape is shifting beneath our hard-to-see feet.

Markers in support of AI include frontier models quickly becoming more powerful and capable. Leading supply chains in our community are starting to see greater returns on AI investments, particularly as they delve deeper into redesigning end-to-end processes and decision-making.

Current and projected capital spending on AI infrastructure also continues at a rapid pace — dwarfing historical infrastructure buildouts. The U.S. interstate highway system was built over the course of nearly four decades with the equivalent of approximately $600 billion in present-day investment. By comparison, AI data center investment is approaching $1 trillion in just six years, and spending is projected to reach $7 trillion by the end of the decade.

Yet, counter to this momentum, there is growing social and political push back on the AI data center build out and significant concerns about AI safety, recently echoed by the leaders of many leading AI companies. At the ground level, my colleagues are taking daily calls from CSCOs and CPOs and hearing endless variations of the rather disruptive request: “Tell me how I can cut 20/30/40 percent of my workforce using AI.”

I’d like to spend the rest of this blog post exploring the issue of trust in AI. A lot of the uncertainty surrounding AI’s future stems from a lack of clarity on whether it will be a net positive for us — individually and collectively.

There are many dimensions to this issue. In the macro, it’s whether AI is a net social and economic good. At a personal level, it is whether AI is going to take your job or if you feel confident in its judgment and discernment in making business decisions.

Objectively, AI has shown tremendous potential in distilling large data sets into summary findings and recommended actions — the basis of better, faster decision making. There is a much larger conversation on whether macro investments in AI will yield positive ROI for the industry or if the myriad impacts of its installation and adoption are a net benefit to society. Those stories will only become clearer several years from now, so I will focus on recommendations at an organizational and personal level.

Based on conversations with, and observations of, our COO community, here are some suggestions on building greater trust around the use of AI in supply chain operations.

• Define and share the “Why of AI.” An organization pursuing AI — or any transformational technology, for that matter — needs to set a North Star direction on how it will improve the performance of its business. To my comment about the near-term, obsessive focus we’re seeing on productivity, this can’t be the sole engine of an AI program. The direction also needs to be about mitigating supply chain risks and unlocking growth opportunities for the business. Realistic targets should be set for any headcount reductions. People will still need to be “in the loop” for important decisions and there are new roles that likewise will emerge in the AI era.
• Focus on uniquely human traits. Most of the leading companies we speak to aren’t planning to shift to full AI-based decision making in their supply chains. It might be a longer-term target of 80% AI with humans in the loop for the remaining 20%. Just as human discernment is still valued for complex decisions, there are other skills and behaviors that need a human touch. Managing strategic relationships with customers and suppliers requires empathy, connection and adaptability that is better suited to people than machines. These roles are adapting to incorporate greater machine intelligence, but they are still held by people.
• Ensuring that decision making is adapted to the times. Leading AI adopters, leverage “bake-offs” to ensure AI-based systems can yield KPI performance that is equivalent to, or better than, more manual-based decision making, before expecting people to shift. Business conditions are always changing, and lately quite dramatically. The algorithms underpinning AI-based decision making will also need to be monitored and updated on an ongoing basis, bringing human judgment to bear. This reality should be factored into any productivity-based resizing of an organization.

Just as it was hard to predict the eventual impact of the internet on business and society a quarter century ago, the same is true for AI today. It is ultimately a more powerful tool whose use depends on the people using and governing it. It is my sincere hope that when the fog lifts, we’ve driven a net positive role for this breakthrough technology.

Stan Aronow
VP Distinguished Advisor
Gartner Supply Chain
Stan.Aronow@gartner.com