Q&A: Genpact Supply Chain Lead on Using AI's Full Impact

As global markets become increasingly complex, Genpact is positioning itself at the forefront of supply chain innovation. 

Tanguy Caillet serves as Genpact’s Global Supply Chain Lead, leveraging more than 20 years of expertise in technology and market dynamics to drive client growth. He oversees an end-to-end practice, from sourcing to after-sales, specialising in integrating Agentic AI with human operations.

By acting as both consultant and operator for giants like Unilever, Tanguy delivers measurable outcomes for 800+ global clients. This unique perspective ensures Genpact’s technology solutions are as practical as they are transformative for the modern enterprise.

What defines a “supply chain decision platform”?

One of the most exciting shifts I’ve witnessed across supply chain operations is moving the discussion from improving processes to making decisions that scan across numerous processes. 

Supply chain decision platforms are emerging as a means to provide individuals who must make decisions (such as plans, KPIs, events, etc.) with every piece of information they need. This requires interconnected systems with a unified data model that ensures the operator has all company data at their disposal to come to an informed decision. 

One of the key components of supply chain decision platforms is that they remove business silos. But what’s even more impressive is that they work at both the horizontal level – integrating processes across order management and logistics, or sourcing and planning – as well as vertically, providing additional granularity and a time horizon for how decisions will play out in the future.

By integrating data across processes and linear time, supply chains also become prime for agentic AI applications, where recurring basic decisions can be intelligently automated with humans-in-the-loop when necessary.

Moving from simply building a resilient operation to generating real-time insights that can power supply chain decision-making is the key theme I’m watching in 2026 and beyond.

What constraints stalled agentic AI in 2025 and what changes unlock execution at scale next year?

The main constraint is data, data, and more data. It sounds cliched because everyone is talking about data, but the lack of quality data that can integrate with AI is a real concern in the industry.

A new Salesforce survey of 6,000 executives found that 84% of respondents have data stacks that won’t work with AI. However, 63% of this same group claim their business is data-driven. There’s a massive gap between where data is now and where it needs to be to enable full scalability of agentic AI and supply chain decision platforms, which require a unified data model, master data, transactional data, and planning data to allow AI to work across silos.

In my opinion, the primary message for supply chain leaders is that there is no artificial intelligence without process intelligence. To unlock execution at scale next year, business leaders will need to redesign their operating models to take full advantage of the specific processes agentic AI can take over and deliver results at scale.

Which high-impact decisions will be automated first across procurement, logistics, warehousing and manufacturing?

I expect the movement of actions will flow from the easiest to the hardest decisions to make, with execution-related decisions the first to be automated.

In most cases, high-impact decisions are usually viewed through a dollar amount. I don’t think that should be the lens through which we need to see it, though. For example, a business wouldn’t decide to automate the decision-making process to build a new plant for a packaging line, even though it's a big financial decision. I believe high-impact organizational decisions are the low-value, high-volume decisions that can be automated. 

The hardest decisions to automate are those that cut across supply chain processes, like a sales order confirmation that requires alignment with logistics carrier availability and production line availability. This type of decision will touch three different systems (at least), and many companies today still simply use manual labour and Excel to perform such tasks. 

What guardrails, audit trails and human-in-the-loop checks keep automated decisions safe and compliant?

Automating decisions requires a comprehensive plan that evaluates speed to value, which can be accomplished by examining data for availability, quality and governance. As its core, better data leads to better AI performance.

Especially where sensitive data is involved, it’s critical to provide safety guardrails whenever decisions are automated. For example, supply chain leaders should factor in potential regulations and policies that could impact the outcomes of automated decisions. Human oversight should also be a standard procedure for monitoring AI actions, which can help prevent errors and ensure the business remains compliant.

All supply chain, sustainability, Scope 3 and net zero leaders should attend:

• Procurement and Supply Chain LIVE: The Net Zero Summit - QEII Centre, London, March 4-5
• Procurement and Supply Chain LIVE: The US Summit - Navy Pier, Chicago, April 21-22

Co-located with Sustainability LIVE, these events brings together CSCOs, CSOs and senior decision-makers at a moment when sustainability, supply chains and commercial performance are increasingly interconnected.

Tickets can be booked online today for The Net Zero Summit and The US Summit. Group discounts available.

How should leaders measure ROI and time to value?

To measure and how not to measure, that is the question. This current wave of AI adoption and integration reminds me of the data science kick in the early 2010s, when every organization seemed to be jumping on the proof of concept.

In that case, the same issue of determining its value arose.

However, every company today has data science teams supporting its decision-making processes with data. It took time to adopt and prove value, but in the end, it did.

One of the lessons we learned from that wave is that organizations need to anchor themselves in what they are trying to do: make better, faster, and more efficient decisions to run complex, global businesses. 

That's what it takes for companies to navigate today’s ever-changing landscape, where disruption and fast-evolving consumer tastes are considered normal.

It is no longer about simply building resilience for disruptions ranging from tariffs to shutdowns, but more about creating adaptable supply chains that can navigate any situation. The way to do this is by making decisions early and with foresight, so the company can pivot if it forecasts that the decision was not an appropriate one. ROI and time to value should be measured in terms of non-value add activities that are removed. 

Today, best-in-class companies are running plans that are approximately 80% touchless, trusting the outcomes of their solvers to free up valuable time for real business decisions rather than number-fudging. This is where companies will see a concrete ROI. The focus should be on automating the most commoditized decisions first to the more complex ones last, eliminating manual processes over time to allow the data to shine.