"AI in enterprises isn’t primarily about the value of replacing human labor. It’s about making human work much more valuable, often in new ways."

About Edmund Zagorin

Edmund Zagorin is co-founder and chief strategy officer of Arkestro, the leading predictive procurement platform accelerating enterprise scale. He is the author of The Predictive Enterprise: How Superhuman Supply Chains Transform Global Business, which examines how AI can reshape enterprise decision-making, procurement and supply chains.

What made you write The Predictive Enterprise, and who did you write it for?

I wrote “The Predictive Enterprise” for executives and procurement and supply chain practitioners who want to advance their careers and deliver wins for their business. The thesis of the book is that in order to be well-positioned over the next decade of AI transformation, you don’t need to deeply understand how AI works. Instead, you need to understand the legacy bottlenecks in your organization’s “process stack”, which is like the interstitium between all the roles and goals of large enterprises. That’s where the big wins are waiting to be uncovered.

To be blunt, a lot of the writing about AI since the ChatGPT moment in 2022 has veered into the traps of irrational exuberance or knee-jerk pessimism. Seeing global AI deployments in enterprise supply chains firsthand, I knew most business readers would want something a bit more brass tacks, detailed, practical, and case study-driven. Not just the “what”, but the “how.”

The other perspective I wanted to move beyond is the tendency to conflate the benefits of AI with the benefits of automation. The stories of the book are about how powerful the Return on Investment from new ways of working can be. AI in enterprises isn’t primarily about the value of replacing human labor. It’s about making human work much more valuable, often in new ways.

What does “predictive” mean in practice for a large company? Could you share one example?

“Predictive” means anticipating the most likely outcome of a process and then changing the process to improve the outcome. Whereas most processes are one-size-fits-all and favor standardization and uniformity, predictive processes with AI can be dynamic and customized.

Think about how Google Maps works. You give it a destination and it compares possible routes using information about where you are and what is happening around you. As conditions change, the route updates. A predictive process should have a similar ability to recommend the best next step based on the outcome you want combined with relevant context. That’s a very different theory of change from purely looking at AI through the lens of automation.

One example involves an electrical-components manufacturer whose existing process made multiple rounds of negotiation too labor-intensive to run consistently. Using predictive procurement where the buyer makes the offer to the supplier instead of vice versa, the team negotiated with 20 supplier organizations in half the time of its previous dealmaking process.

The manufacturer kept the supplier relationship and the same requirements while improving both cost and speed. The prediction became valuable because it changed both how the negotiation happened by re-sequencing the steps. It shows how process architecture can change two-sided incentives that determine cycle time. Whereas reactive negotiations incent mutual delay, predictive negotiations offer a two-sided incentive to commit the deal faster.

Why do so many digital transformation projects fail to deliver, and what should leaders do differently with AI?

Many digital transformation projects require an extraordinary number of things to all go right at once. Employees must change their behavior, systems have to connect, data has to be usable, and several functions have to agree on what success looks like. Each dependency creates another opportunity for the project to stall or regress. Meanwhile, everyone involved still has to do their day job.

For successful AI adoption, leaders should first align on the desired outcomes, then identify the steps needed to achieve them. In procurement, finance might prioritize savings while a plant manager wants to preserve a trusted supplier relationship. Mapping these persona conflicts can allow transformation teams to be clear-eyed about potential failure modes well ahead of time.

Pilots can be a great way to build conviction. Leaders also need to decide what happens after a successful pilot. Who has authority to expand it? What resources will be available? Which results justify moving forward? Those questions should be clearly defined at the start.

Finally, take the additional work of transformation seriously. People already managing a full workload need time, support and recognition for helping drive change. An engaged executive sponsor who clears obstacles and participates in regular project discussions can make an enormous difference.

Who should own AI in procurement and supply chain: the CFO, the head of procurement or IT?

The executive team should own the commitment to transformation and its business results. For initiatives that materially affect costs, working capital or operating margins, the CFO has an important role in setting expectations, sponsoring the work, and demanding urgency.

Procurement and supply chain leaders should drive the changes to their functions. They understand the supplier relationships, operational constraints, and decisions that determine whether an AI use case will succeed. Their teams also need to help define success, because they will be responsible for making that new “process stack” work every day.

IT plays a role as well. However, for many large companies, their suppliers are adopting AI faster. That puts procurement teams on the hook for whether or not suppliers use AI responsibly. The risks from uncontrolled AI are significant.

Procurement can address AI use during supplier selection, contract negotiations and ongoing reviews. The carrot of getting awarded a contract is a very effective way to incentivize AI safeguards that can benefit the enterprise while reducing future existential risks from a potential superintelligence explosion. It’s wild to imagine that procurement’s AI governance may be what ends up saving us all. Yet that’s exactly what my research about long-tail risk events suggests.

If an executive wanted to start this month, what is the first step you would recommend?

Bring together a small, empowered team and identify one recurring problem that people can describe and measure clearly. Include someone who does the job, a budget owner who experiences the consequences and an executive sponsor who can help remove obstacles.

I would start by asking a few simple questions. What slows you down? Where do surprises cost us money? What are our “predictable surprises” – moments where information lives somewhere in our organization but the right person finds out too late? Suppliers should have a voice in this discussion, too, because changes to procurement processes affect their workload.

Then establish a baseline. If the problem is a slow approval process, measure how long it takes today and where the waiting happens. If it is inconsistent pricing, identify the relevant transactions and validate that you are comparing equivalent requirements and contexts.

That gives you a concrete basis for choosing an AI use case and judging whether or not it helps.

In the book, Len DeCandia’s First 100 days playbook emphasizes building this foundation of trust and measurable progress. An early project should be manageable enough to show results quickly, with a clear decision about how it can expand if it works.

The goal for the first month is to understand the problem, agree on the desired outcome and give the right people authority to act.

Get the Book

As an Amazon Associate, The Executive Insight earns from qualifying purchases.

The Predictive Enterprise: How Superhuman Supply Chains Transform Global Business by Edmund Zagorin is available on Amazon.

Latest From The Executive Insight