eCommerce
AI: Physical benefits in a manufacturers world
John Cowie discovered how a company rooted in producing physical goods is embracing artificial intelligence in its operations
After nearly a century of success in paper manufacturing, DS Smith is turning its attention to the next frontier: generative AI. The decision to invest in cutting-edge digital capabilities may seem surprising for a company rooted in physical goods, but for Adrian Pinder, Head of Digital and Data, it represents a logical evolution. “We’ve been making paper very successfully for about 90 years,” he says. “But GenAI is how we think we’ll remain successful for the next 100.”
In a presentation outlining the company’s journey, Pinder described a digital transformation that is at once pragmatic and ambitious—a process that, while it appears coherent in hindsight, unfolded through experimentation, governance, and growing organisational confidence.
Until recently a constituent of the FTSE 100, DS Smith is a £28 billion revenue packaging business, now owned by US-based International Paper. Its operations span 400 factories and 70,000 employees across Europe and North America, producing everything from raw paper materials to sustainable packaging for FMCG giants like Amazon.
Despite its scale and longevity, DS Smith faces a dual challenge: customers increasingly demand innovative, low-carbon solutions, while its IT estate, fragmented by years of acquisition, hinders efficiency. According to Pinder, generative AI offers a way to tackle both issues—enhancing competitiveness for customers and creating coherence across internal systems.
The case for GenAI: The journey began with a call from the board. Curious about the implications of GenAI, senior executives requested a briefing. That initial conversation—grounded in technology fundamentals, risk, and opportunity—gave Pinder’s team a “license to experiment.” While funding was limited, executive endorsement proved more valuable in enabling momentum.
The company’s first move was to put guardrails in place. A cross-functional team spanning HR, legal, innovation, and digital created a six-page Responsible AI Policy. It deliberately excluded HR-related AI use and imposed strict governance, including a controversial but effective block on external ChatGPT usage—helping mitigate shadow IT and underlining the seriousness of the initiative. Next came experimentation. DS Smith developed three initial tools: a secure internal chatbot built on AWS Bedrock with Claude; a translation tool to cut reliance on expensive external services; and a knowledge management system using Retrieval-Augmented Generation (RAG) methods. All three were designed to offer tangible, immediate value. In one instance, the internal chatbot was credited with saving employees 30–60 minutes per day—a small efficiency per user, but significant at scale.
With early pilots yielding promising results, the company undertook a strategic mapping exercise using a Gartner framework to identify broader opportunities. Workshops across business units helped surface dozens of use cases. Importantly, these sessions built on growing internal familiarity with GenAI, thanks to the initial pilots—a sequencing Pinder believes was essential for buy-in.
Executive backing followed, with a clear direction: prioritise back-office automation and core capabilities such as pricing and product design. This clarity enabled DS Smith to shift from experimentation to rollout.
Among the initiatives now being scaled are intelligent document processing for scanned invoices, procurement data analytics layers powered by natural language queries, and continued development of the RAG-based knowledge platform. The broader aim, says Pinder, is to provide every employee with their own “personal intern”—a GenAI assistant capable of enhancing productivity, improving decision-making, and reducing operational friction.
Looking back, Pinder identifies three core lessons for large enterprises seeking to embrace GenAI. First, ambition must be matched with pragmatism. “You have to think big to get people excited,” he notes, “but you’ve got to start small.” DS Smith’s modest initial funding turned out to be a benefit, forcing creativity and focus.
Second, alignment with business priorities is non-negotiable. The digital team framed GenAI not as a standalone strategy but as an enabler—helping divisions reach their goals faster, more cost-effectively, and with lower risk.
Finally, transformation cannot happen in isolation. DS Smith has supplemented a lean internal team with a network of partners—both large and small—with the latter often providing sharper innovation. Strategic hires have bolstered in-house capability, but the company is realistic about the scale of change required. What began as a cautious inquiry from the board has become a company-wide transformation, deeply embedded in DS Smith’s strategy for the century ahead. While the packaging may still be made of paper, the future of the business is increasingly powered by data—and shaped by artificial intelligence.
