Does it resemble hip-hop? Jazz? A cha-cha? The slide? No one knows, but in my discussions, business leaders agree that Artificial Intelligence (AI) is here, and the projected impact on supply chain technology looms large sending the industry into very different gyrations. How fast will it transform work, no one knows, but excitement reigns.

Traditionally, supply chain technology providers evolve slowly, and business leaders move at an even-slower pace (glacier speed) to revise work practices. For me, it is like watching paint drying in the winter. Can we supercharge this? I think maybe.

While teams wave their hands and talk digital, I find digital projects ideas lack definition, clarity of process, and success. (Our research shows that the projects failed to deliver value in 75% of companies.) The industry first started speaking of digital in 2012, but today, it is still an amorphous concept. (My first digital supply chain presentation was in Milan at an SAP Insider conference in 2012.) If we look across the decade, the greatest value happened when business leaders took baby steps in visual analytics, and sadly called it digital.

So where will the innovation come from? New start-ups and business innovators will spark new projects and ideas. Initially, it will be slow (like an awkward waltz). The pace will then accelerate. Perhaps a boogie?

What Is AI? Why Does It Matter?

Let’s start with a clear definition. I define Artificial intelligence (AI) as computing ability to perform human tasks that require human intelligence and discernment. Machine learning and AI are different: machine learning is a way to teach the computer. AI and machine learning are not the same. Machine learning is a means to an end.

The road forward is bumpy. AI outputs, like human thinking, is often flawed. Within an organization, the computer thinking can be consumed with political bias. How wrong and how biased depends on the inputs and the refinement of the model. Controlling error and bias is important because of the speed of computing. Buckle your seatbelt: AI will put supply chain decision-making on steroids. The problem is helping models sort through inaccuracy and bias.

The general AI models like ChatGPT are the buzz, but the greatest lift for the supply chain is happening in the world of narrow AI driven by deep learning. Narrow AI is enabling insights based on disparate data. My remorse is the technologists and business leaders are replacing traditional optimization with newer models based on narrow AI, but not questioning the taxonomies or the current definitions of planning. My analogy is putting a new engine in an old jalopy.

I am pleased at the investment in semantic data models, or ontologies, to compliment graph deployments. Using a knowledge graph (versus a relational database model), companies can analyze shifts in relationship webs, instead of seeing data in separate tables. The knowledge graph enables the visualization of connections between data points that would otherwise be unable to be visualized.

We are a long way from super intelligence where machines are self-aware.

My friends in Silicon Valley working with Venture Capitalists are all a-buzz with use cases. The problem for me, and many of my readers, is that the work on supply chain use cases is far and few between. The fear is the readiness of adoption of a laggard industry where only 3% of manufacturers are early adopters.

What Will Be the Value?

If we are open to the opportunities, and are willing to admit that the current supply chain planning taxonomies and the use of data are legacy, companies make the first step to drive value in in five areas:

• Sense and Respond. Today, companies are investing in networks to improve sensing. Unfortunately, the companies find that they are drowning in data and low on insights. The problem is the lack of a semantic layer. Traditional technology deployments of ERP and APS increase the bullwhip effect and elongate process latency. Unfortunately, companies spend large bucks to make decisions late without clarity on if it was a good decision. The high cost of inflationary inventories and the need to work down post pandemic working capital is a driver for companies to better use network data.
• Relationship Management. Today, companies are blind to relationship flows in their networks. The primary investments are in networks of limited capabilities based on EDI, and indirect procurement spend management. The focus on inside-out processes is a barrier to building strong relationships. The investments in new approaches and deploying outside-in processes enables quicker sensing and bi-directional orchestration across source, make, and deliver. Relationship visibility is also needed in the management of ESG plans.
• Redefinition of Work in Planning. Planning Democratization. I spoke to a company last week with over 700 planners. (Say that again, 700 planners? Yowza, how did this ever happen?) The company defined planning as time to react within the short-term horizons 3-12 weeks. In their zeal to work hard and react, they lose insights on the longer-term choices. Reactivity does not drive supply chain excellence. The goal should be sense and respond. The company asked me, “How many planners should they have?” My answer was, “Are you willing to redefine work?” I believe that the evolution of better engines, architectures, and outside-in processes have the potential to reduce the number of planners by 85-90%. To accomplish this goal, companies need to get clear on the role of planning, the definition of supply chain excellence, and democratize planning. However, if this organization has the current organization deploy newer technology approaches, they will fail. The reason? The planners will fight the redefinition of work.

How Do I Prepare My Team?

The starting point of this journey is to clearly define supply chain excellence. Functional excellence throws the supply chain out of balance and increases waste and decreases margin. Traditional planning taxonomies focus on the optimization of functional metrics. The first step is to define supply chain excellence cross-functionally and align the engines and taxonomies to improve balance sheet effectiveness. Improving cost does not necessarily improve margin.

The second step is to drive organizational learning on new technologies and the Art of the Possible. Stop the stupid RFPs that are circulating in the industry. Instead, partner with a few technologists to test and learn. Do not bound this work by forcing the project to have a defined ROI.

The third step is to map a future plan on outside-in processes and the use of market (both channel and supplier data) to drive bi-directional orchestration end-to-end. There are no technologies in the market yet, to enable the vision, but the interest is high and the technologies more ready to drive the use cases. The barrier is the unlearning. In August, I will be taking three cohorts through a virtual class where we will define an outside-in process, categorize data, align technologies while defining potential use cases/ vision. I look forward to sharing more about the insights of 140 virtual students at the end of August.

Conclusion

These are my thoughts on a hot summer day as I recover from surgery. Two more weeks until I can travel. Stay well and thanks for reading my blog.

I look forward to reading yours. What do you think of the topic of supply chain and AI? How do you envision the dance?