The Cow, The Code And The Chaos: Why Logistics Needs Hybrid Intelligence

Forbes

3 days ago

Shekar Natarajan is the founder and CEO of Orchestro.AI.
Some time ago, I was a passenger in a vehicle in rural India—a place where there is no rule of law regarding driving.
Drivers didn't stay in neat lanes; some ended up on the side of the curb. Drivers could simply raise their right hand to signal a move to the left. They could raise both hands to signal they'd be driving straight. Somehow, in the cacophony and chaos, my driver was able to understand the intent of the other drivers and weave through them.
But then appeared on the road a living creature whose intent my driver could not understand: a cow.
Neither could the cow understand the intent of my driver. Yet in that moment, both had to coexist on that road.
Coexistence, I believe, is necessary for success in all areas of our lives—rather than one or the other, we must find space for both. Technology is not an exception. In my view, for technology to truly help us succeed, we have to approach it through the lens of coexistence. In the supply chain world, AI can help future-proof the industry. In particular, I've observed that GenAI stands to create significant transformations. Yet, for optimal success, GenAI should be complemented by execution engines.
The Limits Of GenAI Alone In The Supply Chain World
Why is solely using GenAI not ideal in the supply chain world?
GenAI behaves probabilistically; it is more predictive and speculative in nature. It exists to simulate, generate and explain. It can take data, uncover patterns and forecast possibilities accordingly.
But the nature of the logistics industry is not probabilistic. Logistics isn't about what could happen—it's about what does happen. GenAI can only go so far in a non-probabilistic reality. For instance, a GenAI solution could analyze past data points and predict that if a retailer ships a package with a certain carrier in a certain city, there is a 90% chance the package will arrive on time. However, that prediction only holds weight in a digital world, which lacks the nuances and disruptions of the physical one. Severe weather, a port blockage, etc., can all cause shipping delays, regardless of the carrier.
GenAI And Execution Engines: The Case For A Hybrid Approach
By contrast, execution engines are deterministic software systems that choreograph and execute concrete actions. They exist to act, decide and choreograph. They revolve around intelligent action, which is the ability to take contextual actions in the physical world. However, acting alone, executive engines lack the flexibility to adapt in real time.
Combined, GenAI and execution engines create a hybrid type of technology called execution engine optimizers. Execution engine optimizers fuse GenAI's reasoning with the capabilities of execution engines to take the most optimal actions in real time. So, GenAI essentially serves as an overlay engine that comprehends intent and simulates. The execution engine does the rest. Execution engine optimizers bind the two worlds together in an intelligent manner, transforming intent into outcomes.
For example, a customer could enter a tracking number into a chatbox, and from there, GenAI in the backend could start its analysis to pinpoint who the consumer is, what their customer lifetime value is and where the package is. By overlaying different data points, such as the weather and traffic conditions, the technology could identify the root cause of the delivery failure (such as a delay or theft). GenAI could then determine the appropriate response to provide the consumer based on the situation. From there, the execution engine could automatically process both a carrier claim and check inventory availability, then offer the customer options such as reshipment or a refund. If the customer, say, opts for a reshipment, the execution engine could create a new shipping order in the system, generate a new tracking ID and send the customer an automated update with the new tracking ID.
In short, these two technologies have different purposes. GenAI processes ambiguous, unstructured data. Execution engines analyze outputs and then execute accordingly. For the best results possible, both of these technologies need to coexist; they need to dance together. Power a supply chain system on GenAI alone, and you're banking on probabilities. Power one on just an execution engine, and you'll get actions without context.
How Supply Chain Leaders Can Leverage The Hybrid Approach
Supply chain leaders can leverage the hybrid technology, called execution engine optimizers, by taking several steps.
First, supply chain leaders should identify their most significant friction points. Those are the problems that, if solved, will yield the most significant net positive results.
Next, I recommend that supply chain leaders use GenAI for analysis and scenario planning. For instance, they could test to see which workflows will get impacted if certain variables change. This information will help them predict potential failures and categorize problems—from there, they should design specific, deterministic processes to address identified issues.
At this point, they can take the outputs GenAI has provided them and start connecting them to execution logic (in other words, automated actions).
As the generative AI engine and execution layer run, supply chain leaders should monitor how the systems are performing and create feedback loops that continuously optimize both the GenAI's and the execution engine's capabilities.
On that road in India, at first, my driver honked and honked for the cow to move. The cow didn't budge. And then he, along with other people, started tapping the cow. The cow looked at everyone, then started walking away. Both parties—people and the cow—were able to create a more stable, streamlined environment. Supply chain leaders have the opportunity to stabilize and streamline supply chains if they put GenAI and execution engines in a symbiotic relationship and take action when needed to keep the duo running together harmoniously.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?