Q&A: Tim Beckhoff, o9 on AI-Driven Demand Planning

In recent years, companies with global supply chains have been forced to undergo constant reconfiguration, driven by geopolitical upheaval. 

As a result, leaders have been turning to AI-driven demand planning to help them better adapt to changing tariffs and trade agreements. 

Through this, organisations can see reduced decision lag, better manage disruptions and build more resilience throughout their supply chains. 

Supply Chain Digital spoke to Tim Beckhoff, Senior Director of Industry Solutions at o9 Solutions to discuss how AI-driven demand planning capabilities are helping companies as they navigate evolving tariffs, through scenario planing and real-time forecasting.

Tim started his career in operations as buyer, specialising in supply chain risk. Throughout this, he explored technology partnerships between his previous company, McKinsey and o9. Here, he gained a passion for delivery-focused responsibilities.

He has now been at o9 for two years, working to deploy concepts into the field.

How have tariffs impacted the global supply chain ability to plan?

Tariffs have added significant uncertainty to plans, making what was previously fairly stable much more volatile. Tariffs have always been part of the plans, but it was more like a long-term configuration that would change over much longer time horizons.

So supply chains had more time to adapt to those changes, making the system more stable. In the actual planning processes, tariffs were not a major challenge. But now what I think is new is using tariffs as a geopolitical instrument, so supply chain planning is getting harder as tariffs become more volatile.

Additionally, there's a need to replan more things in the short term to avoid severe costs, but this requires greater planning agility.

Supply chain planners need to be able to adjust their plans more quickly and predict how a tariff change would affect their supply chain.

They have to plan for scenarios and think about which circumstances give them the option to change the routing, where they buy from and at what point it is then fixed, and they are forced to eat the tariffs. You need to get more agile and the period for which you can plan with some certainty is getting ever smaller.

How are AI-driven demand planning capabilities helping companies navigate tariffs?

Demand is not the first thing impacted by the complexity of tariffs. They affect the supply before demand. Tariffs change the cost structure and then they change the supply, which ultimately shapes demand when costs are passed on to customers.

It's in an incredibly complex system that has many intersections and to create the right scenarios is above what the human mind can comprehend. Even if you take a really smart analyst and build a supply chain model to play that through, the analyst would be stuck with this task for weeks coming up with a good answer.

To understand how this whole chain – from tariff to supply to cost to customer demand – works, I think the most effective tool to really address that is AI-driven planning capabilities and to just have agents that create the scenarios, and play that through simulation, then translate it back into different types of scenarios that then can be assessed from a business management perspective.

What do companies risk by not adopting AI-driven demand planning capabilities?

I think the speed of planning, or the ability to replan with agility, is becoming a stronger competitive advantage because, if supply chains are constantly under threat from disruptions and tariffs, the capability to deliver becomes ever more important.

A great example was the semiconductor crisis, where car OEMs that could deliver sold way more cars. They could effectively capture market share from competitors because of their ability to deliver, regardless of the product's appearance or customer preferences.

It was about supply chain resilience and agility, which were gaining market share. So I think it's losing a competitive edge if a company falls behind, and the competition is more agile and more resilient in planning.

How can o9 help companies stay afloat during turbulence?

o9 can act as the central planning brain that takes all those different signals from demand, from supply and from external risk, then pairs them with a supply chain digital twin and comes up with scenarios nearly immediately after something happened.

Planning with o9 provides way more optionality. If there is a disruption happening and planners need to reroute deliveries, maybe through a different port or use different suppliers, if your company is the first to act, there are still options, and there still might be spare capacity that you can get.

But if you're acting a few days later, once the disruption has hit, you're already too late. So when you only start replanning or set up a task force to help overcome those challenges, and then get into action weeks after those disruptions hit, there aren't many options left. Speed is of the essence, and that's where o9 can really help.

Can you share a client-based case study where o9 has aided in demand planning?

That’s a great question, and to answer it accurately, I think it’s important to first reframe how we look at demand planning.

There is obviously immense value in sharpening your forecast, but treating demand planning as an isolated victory is a legacy mindset. The real magic happens when you leverage a unified architecture like the o9 Digital Brain to let your planning work end-to-end, all the way from the initial demand signal down to supply planning and actual field execution.

If you fix the forecast in a vacuum without natively linking it to real-world operational constraints, you’ve essentially just built a much more accurate map to a dead end.

A phenomenal example of this end-to-end philosophy in action is our work with a major global automotive supplier based in Europe. They were constantly at the mercy of wildly erratic EDI signals from car manufacturers who shifted production volumes on a whim to dodge market uncertainty.

Historically, if an OEM changed their short-term demand, it took our client two full weeks, two entire planning cycles, just to figure out what that meant for their factories and component orders. They were blind in the short term and spending cash on emergency logistics just to keep up.

By introducing the o9 Digital Brain, we completely bridged that gap. The platform used predictive models to read between the lines and decode those noisy, defensive customer signals, which allowed our client to realise a six to 15 percentage point improvement in forecast accuracy compared to their raw EDI data.

But here is where the end-to-end philosophy proves itself: that sharper forecast wasn't just left sitting in a silo to look pretty on a spreadsheet. Because the o9 Digital Brain natively links demand to a fully constrained master supply plan, those clean demand signals instantly collided with real-world material and factory capacity constraints.

That immediate translation into execution completely transformed their bottom line, and because they stopped playing catch-up, the client saw a 30% reduction in express shipment costs and a 6% improvement in global inventory turns.

Ultimately, you don’t fix a demand plan just to get a better number; you do it so your procurement, factories and logistics networks can execute flawlessly.

What do you think the future of supply chain strategy looks like?

I believe the future of supply chain strategy requires a complete departure from traditional, siloed operating models and slow, periodic planning cycles. They simply can't keep up anymore. Instead, the future belongs to the APEX model, which stands for Agile, Adaptive and Autonomous Planning and Execution.

First, supply chains must be Agile. This goes far beyond fast reporting; it’s about detecting subtle shifts, like a sudden demand spike or a port bottleneck, across thousands of nodes in real time. True agility means diagnosing what actually matters and coordinating an immediate, cross-functional response while you still have time to change the outcome.

Second, the strategy must be Adaptive. Instead of waiting for massive technology overhauls every decade, an adaptive supply chain continuously learns from its own history. It systematically analyses plan-versus-actual deviations to identify exactly where value leaked, instantly refining its rules and playbooks for the very next cycle.

Finally, the ultimate North Star is Autonomous execution. Human bandwidth shouldn't be the bottleneck for every tactical choice. Routine, high-frequency decisions, like daily store replenishment or standard inventory deployment, should run on autopilot within strict corporate guardrails.

This doesn't replace human judgment; it elevates it, freeing leaders to focus on high-stakes strategic choices and complex trade-offs.

To make this a reality, future technology must bridge advanced predictive AI with the hard, logical rules of your enterprise footprint.

We need systems that don't just guess, but traceably answer what happened, why it happened, and exactly what the business should do next to maximise value.