PSC LIVE: Sustainability – Capgemini Invent Keynote
At Procurement & Supply Chain LIVE: Sustainability, Nitin Dsouza, Vice President Supply Chain and Engineering at Capgemini Invent, delivered a rich and fast-paced keynote on how generative AI – and more specifically agentic technology – is poised to reshape global supply chains.
Nitin combined technical insight and strategic foresight with a healthy dose of realism about the capabilities and challenges of next-gen AI in the enterprise.
Defining agentic tech
To begin, Nitin drew clear lines between generative AI, agents and traditional supply chain solutions. While ChatGPT and other tools have triggered a wave of excitement, Nitin warned against overhyping their immediate value in supply chain operations.
“You can’t just ask ChatGPT to give you supply chain value,” he said. “You need software around it—agents that solve specific business problems.”
Agentic tech, as he defined it, comprises AI models embedded in software frameworks capable of reasoning, tool use, planning and reflection. This makes them fundamentally different from standalone language models.
The value, he argued, lies not in replacing existing tools but in complementing them through layered automation, smart interfaces and workflow optimisation.
Agentic use cases currently cluster in low-risk, high-volume activities: email parsing, research, report generation, code configuration and documentation.
“Wherever there’s text—there’s opportunity,” he said, noting that procurement, training and documentation workflows are prime targets.
The supply chain evolution
Nitin traced the history of supply chain development across three key eras: lean and agile systems in the 1980s–2000s; digitally connected chains in the 2000s–2020s; and now the potential dawn of the agentic-powered supply chain.
Yet, for all the change, he insisted that core business imperatives remain constant: lower cost, faster time to market, agility, sustainability and resilience.
“We’re always in perma-crisis,” he quipped, referencing geopolitical volatility, shifting tech stacks and evolving production models.
Drawing comparisons between Toyota’s lean supply principles and Tesla’s real-time, software-integrated production systems, Nitin illustrated how digitisation is beginning to reshape physical operations. But he also made clear that not all aspects of AI are ready to replace legacy solutions.
“Existing planning and optimisation algorithms still outperform generative models in many areas,” he added.
This includes forecasting, orchestration and production planning – areas where mature AI or even conventional systems still have the edge due to accuracy, speed and reliability.
Navigating architecture, UX and environmental impact
Nitin took time to walk the audience through the emerging AI stack: from cloud infrastructure and GPU acceleration to orchestration layers and human-agentic interfaces. In supply chain contexts, he sees Gen AI operating as a decision layer above existing analytics, triggering tasks and providing insights – but not replacing critical functions.
He discussed evolving user experiences, particularly the difference between interfaces designed for humans and those optimised for agents. As more tasks are handled machine-to-machine, UX design is beginning to shift.
Environmental impact was another key area of discussion, prompted by an audience question.
“Yes, today’s AI has a significant carbon footprint,” Nitin admitted, citing studies that show large language models consume substantial energy.
But he argued that the tech’s long-term potential to support sustainability – by streamlining operations, optimising materials and accelerating innovation – could outweigh the short-term cost.
“I’m on the side of using it where it matters,” he said, noting that LLMs are already being used to compress RFP timelines, surface insights from massive datasets and guide configuration for supply chain systems.
Key takeaways
Throughout the session, Nitin balanced excitement with caution.
“Not every shiny object is gold,” he continued, referencing his earlier work on RFID’s rise and fall. Yet he urged professionals to explore, experience and educate themselves on agentic technology, emphasising that it’s better to be prepared than surprised.
In his closing remarks, Nitin underscored three key takeaways:
- Human and agentic workforces will increasingly operate side by side.
- Agentic technology is in its early stages, but exploring real use cases now is essential.
- Supply chains should augment – not replace – traditional systems with agentic layers.
Asked whether agent-to-agent communication might replace human interaction, he remained pragmatic.
“Why do we need agents to speak to each other in their own language when we have APIs?” he asked. “It’s interesting, but it’s also a bit of hype.”
With deep industry insight and a clear-eyed view of the future, Nitin left the audience with the sense that while agentic technology isn’t a silver bullet, it is a powerful toolkit – one that supply chain leaders should begin exploring today.
Essential diary dates for 2025
Discover the essential diary dates for Procurement Magazine and Supply Chain Digital, as well as its sister publications, Manufacturing Digital and Sustainability Magazine.
To follow Procurement & Supply Chain LIVE on LinkedIn, click here.
To enter for the Global Procurement & Supply Chain Awards, click here.
- Sustainability LIVE Chicago | 28-29 May
- Procurement LIVE Chicago | 28-29 May
- Supply Chain LIVE Chicago | 28-29 May
- Procurement & Supply Chain LIVE London | 23-24 September
- The Global Procurement & Supply Chain Awards | 24 September
Explore the latest edition of Supply Chain Digital Magazine and be part of the conversation at our global conference series, Procurement & Supply Chain LIVE.
Discover all our upcoming events and secure your tickets today.
Supply Chain Digital is a BizClik brand.
