BMW Group's Raúl Gamboa on building agile, accurate and AI-driven vehicle distribution
Speaking at the Finished Vehicle Logistics North America 2026 conference, Raúl Gamboa, head of logistics at BMW Group’s San Luis Potosí Plant, outlined how the company is reshaping its North American distribution strategy through a new cross-border Landbridge Project, a suite of in-house AI tools now live in production, and a digital transformation philosophy built from the plant floor up – before joining a panel alongside Glovis America and ICL to explore what connecting the vehicle supply chain through data actually requires.
On Wednesday 22nd April, on the opening morning of the Finished Vehicle Logistics North America 2026 conference, Raúl Gamboa, head of logistics, production control and production systems at BMW Group’s San Luis Potosí Plant, opened with a framework he uses to categorise the challenges facing current logistics operations. The first layer is global disruption: geopolitical instability, shifts in trade policy and port failures. The second is specific to Mexico: farmers blocking highways and political decisions affecting logistics infrastructure.
“I’ve been listening throughout the conference to many of the disruptions and challenges we face,” he said. "On the global context, which is basically the first layer, many of them we cannot drive or steer ourselves, but certainly something we need to face every single day. On the second layer, in Mexico, we have our own issues – farmers blocking our main highways, or critical decisions from our politicians. This is something we need to live with on a daily basis."
The third layer – the one his keynote was built around – is digital transformation. While tariff uncertainty and market volatility sat in the background of the session, as they have throughout the conference, Gamboa focused on the initiative that sits at the intersection of all three layers: a new approach to how BMW Group gets its vehicles to market.
The Landbridge Project
Gamboa announced the launch of the Landbridge Project – a shift in how BMW will move finished vehicle from San Luis Potosí to the North American market.
Where the plant has historically relied on sea freight for US volumes, the Landbridge Project will route vehicles across the Mexican border by rail or truck, initially serving Texas and California, with scope to expand further as the programme matures.
Gamboa said the programme is expected to cut delivery times by eliminating the time spent transporting vehicles to port, staging in yards, loading onto vessels, and then distributing from arrival ports to vehicle distribution centres.
“With this process, we will be able to cross the border and receive all the vehicles to the BDCs directly,” he explained.
On cost, BMW is targeting a reduction of up to 30% in total outbound logistics expenditure from the operation.
The project was enabled by recent digital systems integration across finance, customs and operations.
Building the digital foundation
That same logic rooted in the Landbridge Project – that sustainable innovation must be rooted in operational discipline – runs through BMW San Luis Potosí's digital transformation programme too. Gamboa outlined three pillars underpinning his team's approach: people, processes and solutions.
On the people front, the plant benefits from a young and highly engaged workforce, with an average employee age of around 30. “You can imagine these young, very motivated people, who are enthusiastic about driving and promoting digital transformation,” Gamboa said. To harness that energy, the plant hosts hackathons and development events throughout the year, welcomes specialists from BMW’s German headquarters to support local engineering teams, and provides structured training programmes to build long-term capability.
The process pillar reflects a measured and value-driven approach to transformation. Gamboa emphasised that BMW Mexico is focused on targeted use cases of AI that delivers clear operational benefits. “In order to make sure that we are profitable in the implementation of AI tools, we have to define first how we want to make this approach,” he said. The strategy begins with identifying repetitive processes where digital tools can create meaningful value, before progressing through governance, implementation and measurable outcomes. “We have been able to proudly say that we have achieved the third layer in these regards, because we have been able to set up processes that have already given us what we were expecting.”
On the solutions side, three key technology investments provide the foundation for this strategy. These include the recent rollout of SAP S/4HANA in Mexico, completed just four weeks before the conference, enabling cloud-based and real-time data access across BMW’s global network; the GAIA AI platform, which allows the plant to manage and protect data sharing with internal and external partners; and a broader cloud infrastructure designed to give all functions consistent, real-time access to the same information.
AI tools with real results
Building on that foundation, Gamboa showcased a portfolio of live, in-production AI and digital tools – all developed in-house by BMW’s Mexican engineering team – that are already delivering measurable operational gains:
- The overseas prediction platform provides real-time visibility of inbound freight, enabling faster decisions on switching transport modes or adjusting production schedules when component supply is at risk.
- Live warehouse process mining delivers automated, real-time tracking of material flow across the plant, helping ensure components reach the assembly line at the right moment and production commitments are maintained.
- The yard integration process – already live for inbound operations and now being extended to outbound flows – is designed to automate driver processing. As Gamboa explained, it works “like an ATM machine”, allowing drivers to scan documents, receive routing instructions automatically, collect the required vehicles and leave the facility with minimal manual intervention.
- The vehicle distribution planning tool: a fully AI-powered, cloud-based system that combines real-time data from logistics partners – including vessel schedules, trucking capacity, rail availability, yard space and finished vehicle readiness – to generate daily shipment plans. “We use AI calculations in order to give us a recommendation on what is the best way to transport our vehicles — whether it should be by sea freight, by rail car or by truck,” Gamboa said. “And the most important thing: all of these decisions are based on the most cost-efficient way to do it.”
Underpinning these capabilities is a control tower providing live visibility across yard, port and transport operations, alongside a cost transparency tool that consolidates outbound expenditure into a single, accessible view, Gamboa explained.
BMW’s transition to the Neue Klasse
Looking ahead, Gamboa said BMW Mexico's goal is to progress from analysis and optimisation towards genuinely autonomous, AI-led decision-making. But reaching that level of automation depends on resolving three challenges: connectivity (ensuring the high-level platform capabilities required by new vehicle architectures are equally accessible in Mexico); availability (keeping pace with the speed of AI development, even with a skilled and motivated workforce); and – most fundamentally – data quality.
"We need to be very sure that the data we are receiving and the data we are creating have enough level of quality in order to be used for this. So these are some of the new challenges that we will have in this digital transformation." – Raúl Gamboa, head of logistics, production control and production systems, BMW Group Plant San Luis Potosí
Those challenges will be compounded – and made more urgent – by the approaching Neue Klasse transition. BMW's recently announced next-generation vehicle architecture, accompanied by the sixth generation of its electric vehicles, will be produced at San Luis Potosí from next year. New logistics opportunities, new data requirements, new complexity – but also, Gamboa suggested, a natural inflection point at which the digital groundwork of recent years will be tested at scale.
"We are looking forward to developing this in collaboration with our service providers and our partners in logistics," he shared.
Delivery through data: Connecting the vehicle supply chain
Immediately following his keynote, Gamboa joined a panel discussion that extended the morning's themes into a wider examination of what data-connected finished vehicle logistics looks like in practice. He was joined on stage by Sean Moses, head of vehicle logistics planning at Glovis America, and Jessica Babajan, director of business development at ICL, for a session that covered everything from the fundamentals of clean data to the emerging challenge of agentic AI – and the people questions that sit behind both.
ETA: where everything starts
Moses opened with a focus on customer satisfaction. "ETA is the highest priority for customer satisfaction – for any new startup OEM, whether it's direct to consumer or a legacy wholesale." Babajan added that the most valuable outcomes for her customers centre on reducing dwell time and surfacing exceptions faster, allowing logistics teams to focus on problem-solving rather than problem-hunting. "Reducing transit time, reducing dwell time – all of those points," she said.
The single source of truth
A recurring theme was the critical importance – and persistent difficulty – of achieving clean, unified data. Moses said: "Every auto manufacturer is growing new silos, more on-premise servers – everything is so fragmented… AI only knows what AI knows. And if you don't have the data set in one place, you can have all the AI projects you want and you won't be ready for it."
BMW’s recent SAP S/4HANA rollout – which went live just four weeks before the conference – has significantly strengthened data validation processes, Gamboa said, creating a more robust and transparent operating environment. As a result, data inputs that may previously have passed through legacy systems now require a higher level of accuracy and consistency.
“As much clean data, as quality of the data, and from the source it’s coming from – that has been quite an effort the last four weeks,” Gamboa said, noting that the transition will continue as BMW works closely with service providers and suppliers to further enhance data quality standards across the network.
Babajan offered practical counsel for those early in the journey: start with step one. "Work with all carriers and OEMs to make sure they're able to accomplish just that first step right – before going to real time. If you have any amount of data but you can't trust it, it becomes irrelevant."
Real time vs. milestones
On the question of real-time location data versus milestone-based tracking, the panel landed on a nuanced view. Moses saw value in milestones as a starting point, but was clear that telematics – already standard across virtually every new vehicle – offers a route to the precision needed for last-mile ETA accuracy. "For inventory reconciliation… let's get that telematics data, reconcile the inventory, and for last mile especially, telematics data is key." Babajan added that real-time data is only valuable if it will be acted upon: "If it's not going to be used properly, there's no reason to get to that level."
Gamboa noted that BMW's control tower in Mexico already draws on GPS, geo-fencing and EDI data from logistics partners – and that one of its most practical recent applications has been monitoring potential highway blockages caused by farmer protests, enabling the team to reroute shipments before disruption escalates.
AI maturity
Babajan described ICL's current work as "dabbling in a little bit of everything" – from internal process automation to truckload optimisation and network capacity forecasting – with agentic AI beginning to enter the picture. Moses said Glovis already deploys an AI chatbot for ETA queries across several customers, and sees the claims process as a particularly strong candidate for agentic AI. "Anything where you can give a clear target and parameters with data to support, you have exponential opportunity for AI," he said – while acknowledging workforce anxieties: "Today, just making every team member more powerful is really the next step."
For Gamboa, the single biggest barrier to more autonomous AI decision-making is accountability. "We still foresee the necessity of a human being to make the last decision. There are very strict policies on what we can share and what cannot." He was equally clear that this is not fundamentally a technology problem: "This is not an IT topic. This is an organisational, operational topic. And everybody has to be responsible for the data they have access to."
On winning hearts and minds
The panel's closing exchanges turned to the human side of technology adoption, and Babajan advised: "Educate them – they don't know what they don't know. If they're educated on the tools, educated on the value it will bring, then they'll trust it, and then they'll use it. Give them the knowledge, then convince them, and adoption will happen from there." Moses suggested engaging people at the level of the problem itself: "Go to your claims team, your port team, whoever is going to use this tool, and say: none of us are perfect. Look at this as another tool. Just like Excel, just like every other tool we've used along the way."
Gamboa reflected on the particular dynamics at BMW Mexico – a young, enthusiastic local engineering team working alongside German colleagues with decades of institutional experience. His overarching advice for navigating any digital transformation journey: "Before going into very high levels of innovation or digital transformation such as predictability, what we encourage first is to have transparency on the information you have. That will be the first step before going into much higher approaches."
