A recap of the Supply & Demand Chain Executive webinar featuring Ruben Munoz-Aguilera, Global Risk Manager at Roche Diagnostics, and Kamal Ahluwalia, CEO of Resilinc.
When 70% of all medical decisions depend on diagnostics, supply chain disruption is not a business problem. It is a patient care problem. That was the framing Ruben Munoz-Aguilera brought to a recent Supply & Demand Chain Executive webinar, and it set the tone for a candid, metrics-rich conversation about how Roche Diagnostics transformed one of the world’s most complex supply chains from reactive to predictive.
Ruben, Global Risk Manager at Roche Diagnostics, joined Kamal Ahluwalia, CEO of Resilinc, to share the full arc of the journey: the scale of the challenge, the hard lessons about visibility and alert noise, and the measurable outcomes that are now reshaping how Roche’s procurement and risk teams operate every day.
The scale of the problem
Roche Diagnostics is not a typical manufacturing company. Founded in 1968, the division now delivers more than 30 billion diagnostic tests annually across 600,000 customers worldwide. It operates 14 specialized manufacturing sites on four continents, three distribution hubs, and more than 11,000 products. In 2024, it reported $14.3 billion in sales.
Behind that scale is a supplier network of 3,250 suppliers across 6,000 sites — and 300-plus internal users spread across a decentralized procurement organization. For Ruben and his team, the job is to protect every link in that chain.
“Seventy percent of medical decisions rely on diagnostics,” Ruben said during the webinar. “Diagnostics functions at the foundational level — it touches safety requirements for people. That translates for me as a persistent call to the responsibility we hold for maintaining continuity, quality, and trust throughout the entire supply chain.“
Why diagnostics supply chains are different
Most supply chain risk conversations focus on cost exposure or delivery delays. In diagnostics, the stakes are different in kind, not just degree.
Ruben walked through the complexity embedded in a single PCR test kit — a product that looks simple from the outside but draws on polymers, plastics, cellulose, active biological reagents, buffer solutions, chemicals, and cold chain packaging, each sourced across multiple tiers. And that’s just a test kit. Roche’s Cobas 6800, a fully automated high-throughput molecular diagnostic system, contains more than 3,000 unique bill-of-materials items. Tracking material flow for a product like that means following the supply chain down to raw molecules and in some cases, to rare earth minerals. Now multiply that across an entire portfolio of instruments, reagents, and consumables, and the visibility challenge becomes almost unfathomable.
“Our visibility needs to get to places where we have concentrations,” Ruben explained. “Where we need to understand: if something happens here, how is this going to affect us downstream?”
Additional pressures compound the challenge: extreme regulatory requirements, cold chain dependencies, demand volatility, limited alternative suppliers for specialized materials, and increasing geopolitical exposure. Unlike automotive or consumer electronics, Roche cannot easily relocate validated suppliers or manufacturing sites. The investments are long-term, the validation requirements are strict, and patient safety is always the ceiling.
Where Roche started: Fragmented data, manual work, alert noise
Before partnering with Resilinc, Roche’s risk management practice was hampered by a familiar set of problems: disconnected data, manual processes, and information that arrived too late to act on.
“We had a lot of information, but mainly disconnected and incomplete pieces,” Ruben said. “Connecting the dots required a lot of manual effort. And I can tell you that was often a frustrating effort. You get where you want to get, but the effort you had to dedicate to it was sometimes really daunting.”
The team was receiving alerts, but the signal-to-noise ratio was working against them. One of the early experiments at Roche told the story clearly: they split users into two groups. One group received alerts only for the specific suppliers they managed. The other received alerts for every supplier in the system. The reaction was immediate. The first group found the alerts useful and actionable. The second felt overwhelmed and disengaged.
“Simple things can make a big difference,” Ruben said. “If you bring the right information to the people who have the right context, you are already making it much better.“
The broader problem was structural. Risk data was siloed across teams responsible for component risk, supplier qualification, compliance monitoring, and scenario modeling. There was no unified view, no clear ownership of action, and no systematic way to understand business impact before a disruption had already arrived.
Building the foundation: Sense, Recommend, Act
Roche’s vision for transformation centered on four strategic drivers: unified risk intelligence, extended multi-tier visibility, scalable supplier engagement, and actionable workflows. Working with Resilinc, the team built the data and intelligence foundation required to pursue each one.
The Resilinc platform provided material breakdown down to the component and sub-component level, tracking products, parts, and materials across the full supply chain. It then combined it with autonomous supplier mapping to extend visibility beyond Tier 1. That structural foundation enabled what Kamal described as the core capability Roche needed: not just sensing disruption but being able to act on it.
“We realized that if we really bubble it up to what we’re solving for, it is material flow and revenue at risk,” Kamal said. “Material breakdown establishes the material flow. When you marry that to supplier information, you have an accurate map. Then when you can sense an issue across multiple vectors, the ‘so what?’ becomes the critical question. Is it a fire drill, or is it good to know but nothing needs to be done because I have enough inventory for that region?”
This is the logic embedded in the Sense → Recommend → Act framework. Resilinc’s platform surfaces real-time disruption signals across more than 40 event categories and contextualizes them against Roche’s specific material exposure and supplier footprint. From there, it generates recommendations aligned to Roche’s own standard operating procedures and business continuity plans. Action can then follow through workflow automation, supplier outreach, or integration with Roche’s enterprise risk decisioning systems.
For a regulated industry like diagnostics, the human-in-the-loop design was intentional. “The human in the loop is essential,” Ruben said. “The system is going to help the person behind it be really engaged. Otherwise, nothing is going to work. This is the key element that makes everything work.”
Confidence at scale: The role of AI-driven scoring
One of the more concrete innovations Kamal described was Resilinc’s confidence scoring model. This is a capability built specifically to address the reality that in global supply chains, you will never have complete, clean data before you need to act.
“There will never be a time, especially in a Fortune 100 company, where you have a complete, clean data set to empower your decisions,” Kamal said. “Models and people and experts all need to know how to work with incomplete, noisy information.“
The confidence score aggregates signals from trade data, third-party intelligence, Tier 1 supplier portal inputs, community mapping, and validated customer data to produce a probabilistic view of a supplier relationship or risk event. It provides the explainability and traceability that allow risk teams like Roche’s to act without having to validate every data point manually.
“The confidence score solves the problem of uncertainty and trustworthiness in automatically discovered supplier relationships,” Kamal explained. “It allows the system and users to make accurate, risk-adjusted decisions without human validation of every data point.”
For escalation decisions, the score provides a calibrated threshold: when confidence is high, action can be automated or expedited. When it is lower, the human in the loop stays engaged. In a regulated industry, that balance matters enormously.
The results: 90% less manual work, 15x more supplier coverage
The numbers Roche shared during the webinar are hard to argue with.
Since implementing Resilinc, the team achieved a 90% reduction in manual effort for risk monitoring and supplier assessment. A process that previously required extensive back-and-forth interactions and months of effort per cycle now runs continuously and at scale. Ruben was direct about what that felt like: “The level of frustration is almost gone.”
Coverage expanded by 15 times, from assessing a few hundred suppliers per year to actively monitoring virtually all direct material suppliers, with deep multi-tier visibility into 150-plus critical suppliers. More than 300 internal users are now active on the platform, each receiving contextualized alerts for their specific supplier portfolios.
“When I started this journey, we were able to assess a couple hundred suppliers per year,” Ruben said. “Now we are capable of actively monitoring all our direct material suppliers. And the people who are using the system, they go in every day, they receive alerts as soon as something happens, and they have the right context. For me, this is the most important thing.”
The cultural shift inside Roche was equally notable. Teams that once felt overwhelmed by fragmented data and reactive processes now have a sense of operating with agency. “What I see is a sense of having things more under a certain level of control,” Ruben said. “My colleagues are a bit more at ease with what happens because they have the tools required to act. As soon as something happens, we can react in a way that we know whether it requires a simple action or a specific task force, and the information is available much faster.”
What’s next for Roche
When asked about the road ahead, Ruben named two priorities: deepening the partnership with Resilinc and building stronger supplier collaboration and workflow integration.
“We can only do this if we develop a deep collaboration with key service providers,” he said. “Our collaboration with Resilinc is an example of that. We need to make sure we drive innovation and capabilities together. A highly collaborative relationship — that’s what we need to keep pushing forward.”
The second priority is internal and external workflow integration, strengthening the flow of work between Roche’s procurement teams and its Tier 1 suppliers so that risk intelligence translates into faster, better-coordinated action across the value chain.
The broader vision is an intelligence-driven supply chain platform: one that integrates external risk signals with internal data, provides genuine multi-tier visibility, and enables scalable supplier engagement at the pace the business requires.
The bottom line
Roche Diagnostics operates in a domain where supply chain failures have human consequences. What Ruben and his team have built, with Resilinc as the intelligence and agentic layer, is a program that treats risk management as a continuous operational capability, not a crisis response.
The combination of material breakdown modeling, autonomous supplier mapping, contextual alerting, and the Sense → Recommend → Act workflow has given Roche something rare: the ability to get in front of disruptions before they affect production, compliance, and ultimately patients.
“If you detect a risk early on,” Ruben said, “you can really prevent many things from happening. If you get the information when it’s already on you, the energy is super difficult to stop without big consequences. Multi-tier visibility is essential if we really want to stop things before we, and our patients, are affected.”
That is the promise of agentic AI applied to supply chain risk. And for one of the world’s most critical diagnostic manufacturers, it is no longer a vision. It is already working.
