IBM and PepsiCo are demonstrating why successful enterprise AI begins long before the technology itself. Through years of transformation, the two businesses have created a connected global platform built around clean data, harmonised processes and an operating model capable of supporting intelligence at scale.
For many organisations, the rapid rise of artificial intelligence has created an urgent question: how quickly can it be deployed? For IBM’s Omar Syed, Lead Client Partner, SAP Leader UK&I at IBM Consulting, the more important question comes first: is the business genuinely ready for it?
In conversation with Business Enquirer journalist Anne Marie Hagerty, Syed explored how IBM’s long-standing relationship with PepsiCo has demonstrated that enterprise AI does not begin with algorithms, agents or automation. Instead, it starts with the less glamorous but far more important foundations of clean data, harmonised processes and a technology environment capable of supporting intelligence at scale.
Through years of transformation, IBM and PepsiCo have worked to turn a complex, fragmented technology estate into a connected global platform, creating the conditions for AI to move beyond isolated pilots and become an increasingly integrated part of everyday business operations.
That journey entered a defining phase in 2019, when PepsiCo embarked on an ambitious programme to modernise systems and processes that had developed market by market and function by function over decades.
“Like most enduring partnerships, it began with a problem that needed solving,” Syed explains. “PepsiCo were looking to modernise their fragmented processes and technology estate that had grown market by market, function by function over decades.”
IBM brought extensive SAP capabilities alongside deep consumer-products expertise, but Syed believes another characteristic proved equally important: a willingness to take responsibility for outcomes rather than simply deliver individual components.
The scale of the challenge was considerable. PepsiCo was seeking to create a globally designed operating model spanning major markets and functions, from planning through to financial reporting.
Standardising at Scale
Rather than beginning with a blank sheet of paper, IBM brought a preconfigured food and beverage solution based on SAP best practices and consumer-industry standards. This provided the programme with a functioning starting point while helping IBM and PepsiCo adopt what Syed describes as a “why not standard?” mentality.
Instead of customising technology around every historic way of working, the organisations challenged themselves to adopt recognised industry-standard processes wherever possible.
That discipline, Syed says, became crucial to delivering a transformation of such scale.
“The discipline of adopting industry-standard processes rather than customising everything is what kept a programme of this scale on time and on budget.”
Building for What Came Next
When the programme began in 2019, artificial intelligence had not yet become the dominant boardroom conversation it is today. The immediate priority was operational excellence: consolidating systems, standardising processes and creating a common operating platform across PepsiCo’s major markets.
In the UK alone, Syed says 13 legacy systems were retired, with a similar number removed in Poland during the programme’s first wave of go-lives.
Crucially, the transformation was not simply imposed on the organisation. PepsiCo assembled around 300 subject-matter experts and 34 global process leaders to help co-design the solution, ensuring that the business itself played a central role in shaping the platform.
That collaboration also created something that would become significantly more valuable several years later: a trusted and harmonised data foundation.
Syed points to a long-held IBM philosophy articulated by CEO Arvind Krishna: AI can only deliver meaningful enterprise value when it sits on top of reliable, connected data.
The standardisation of data models, processes and financial structures was therefore not simply good ERP practice. It was preparation for a future in which intelligent systems would increasingly depend on that information.
“We knew the AI layer was coming at some point,” Syed says. “And we built the foundation to receive it.”
When generative AI and enterprise AI subsequently accelerated, PepsiCo did not have to begin the laborious process of rebuilding its information architecture from scratch. The foundations were already in place. That distinction, Syed argues, is what allows businesses to progress beyond experimentation.
“PepsiCo didn’t have to retrofit all of their data estate,” he says. “It was ready. They were able to move beyond just AI pilots because the data underneath them was actually connected.”
“We knew the AI layer was coming at some point. And we built the foundation to receive it.”
From Operations to Intelligence
Syed describes the IBM and PepsiCo relationship as having evolved through three broad chapters.
The first centred on business operations, with IBM supporting processes and developing detailed institutional knowledge of PepsiCo’s systems.
The second was transformation: reimagining the way the enterprise operated across areas including finance and supply chain.
Now comes the third chapter: intelligence.
The clean data foundation established through the previous transformation is increasingly being converted into AI-enabled capabilities.
It is an evolution that also demonstrates why Syed believes successful technology partnerships depend upon something that cannot be written into a software specification: trust.
One example came when the original global deployment was estimated to take around a decade. PepsiCo challenged IBM and SAP to find a way of achieving it in five years.
For Syed, that was indicative of a relationship that had moved beyond a conventional customer-vendor dynamic. IBM, PepsiCo and SAP returned to the deployment strategy and collaboratively found a new path forward.
Not long after the programme began, COVID-19 fundamentally changed the operating environment. Teams were forced into remote delivery, supply chains experienced extraordinary disruption and employees across different time zones had to continue managing a major live transformation under unprecedented circumstances.
The programme continued. More importantly, Syed says the relationship emerged stronger.
One Global Language
Asked what achievement makes him most proud, Syed points not towards a particular piece of technology but towards the foundation created beneath it.
For a global consumer business such as PepsiCo, future growth increasingly depends on the ability to launch products faster, manage inventory more effectively, react quickly to changing demand and operate consistently across multiple markets.
The platform developed through the transformation supports all four.
Markets can increasingly operate using common data structures, process definitions and key performance indicators, creating what Syed effectively describes as a shared language across the enterprise.
It is this consistency that opens the door to AI applications in areas including demand forecasting, procurement and supply-chain planning.
There is also a human dimension. Once employees trust the underlying data and begin to see intelligent technology improving their work rather than threatening it, Syed believes adoption changes fundamentally. Instead of organisations constantly pushing new tools towards employees, demand begins to come from the workforce itself.
The transformation therefore invested heavily in PepsiCo subject-matter experts and super users from the outset, developing people who could become long-term custodians of the platform rather than temporary contributors to a technology project.
“That’s how you sustain a transformation rather than just deliver it,” Syed says.
From AI That Advises to AI That Acts
With the technological backbone established, the next phase promises to be considerably more intelligent.
Syed sees several major changes emerging over the next five to ten years, beginning with the transition from AI-generated insight to AI-generated action.
Today, intelligent systems predominantly tell people what they should consider doing. Increasingly, agentic AI will be capable of carrying out activities itself within predetermined controls.
For PepsiCo, Syed suggests this could eventually encompass activities such as rebalancing inventory, triggering replenishment or adjusting trade promotions with retailers.
Human decision-makers would remain central, but their role would increasingly focus on establishing objectives, controls and guardrails rather than manually pulling every operational lever.
A second shift will come through personalisation. Consumer-goods companies have traditionally marketed towards demographic or behavioural segments containing thousands or millions of people. AI could increasingly make it possible to engage consumers on an almost individual basis.
Meanwhile, supply chains are likely to become progressively more self-optimising, continuously sensing changes in demand, simulating potential scenarios and replanning accordingly.
Yet Syed believes there is another dimension that receives considerably less attention: people.
“Those changes need people to work differently,” he says. “Not just use new tools, but adopt new behaviours, new decision-making instincts and new ways of collaborating with AI agents.”
IBM consequently treats workforce strategy as a formal workstream within major transformation programmes, mapping how individual business roles could change and identifying the behaviours employees will need in an AI-enabled environment.
For Syed, the organisations that manage this successfully will not merely implement technology on schedule. They will continue creating value from it long after implementation.
The New Constraint
As artificial intelligence becomes more powerful, Syed expects one of the biggest obstacles to adoption to shift.
If intelligent systems begin making increasingly significant operational decisions, companies will need confidence in how those decisions are reached.
Are the models fair? Are they secure? Can their actions be governed? Can the organisation understand and control what they are doing?
For a business serving hundreds of millions of consumers, those questions become particularly significant.
It is here that Syed sees IBM’s role evolving alongside the technology itself: combining enterprise platforms with governance frameworks and decades of experience operating mission-critical systems.
“Our job fundamentally is to help clients go faster safely,” he says.
The principles demonstrated through PepsiCo could consequently extend well beyond the consumer-goods industry.
Businesses separating from parent companies, organisations rebuilding their digital infrastructure and companies in highly regulated industries increasingly face similar challenges around global supply chains, information consistency and AI readiness.
Syed’s prescription is deceptively straightforward: establish a clean data foundation, harmonise processes and only then layer artificial intelligence on top with appropriate governance.
“Our job fundamentally is to help clients go faster safely.”
High Ambition, Low Ego
Technology occupies much of Syed’s professional world, but his approach to leadership is distinctly human.
“My leadership philosophy is quite simple,” he says. “It’s high ambition, low ego.”
For Syed, effective leadership begins with clarity. In a programme involving hundreds of people across multiple countries, employees should never have to guess what success looks like.
The leader’s responsibility is therefore to remain relentlessly clear about the destination while allowing teams flexibility over the route they take to reach it. His second principle is psychological safety around experimentation.
“If people are punished for intelligent failures, you’re going to get compliance and not innovation.”
The rapid shift to remote working during the pandemic demonstrated this particularly clearly. Teams had to rethink complex global delivery almost overnight, something Syed believes could only happen because people felt able to experiment and adapt.
His third principle becomes most important when circumstances deteriorate.
“Leadership is most visible when things aren’t going well,” he says.
When programmes are under pressure, clients are frustrated or teams are exhausted, Syed believes leaders have to be visible, calm and transparent.
“Credibility isn’t built when things are going well. It’s built on how you respond when they aren’t.”
Learning in Public
Working across countries and cultures has provided another important lesson.
Technology itself may increasingly be standardised, but implementing it remains, in Syed’s words, “a deeply human experience”.
Behaviours, expectations and working practices differ significantly between markets, making it important for global organisations to listen before prescribing solutions.
The arrival of AI has demanded another form of adaptability from leaders themselves. Syed describes having to “learn in public”: remaining curious about technologies that did not exist when he began his career and being willing to ask questions that might occasionally sound basic.
Rather than viewing that as a weakness, he considers it one of the most important behaviours modern leaders can demonstrate.
For those beginning their careers, his advice follows three themes: understand the business rather than focusing solely on technology, build a reputation around delivery, and remain curious beyond the point of comfort.
“There are many strategy decks out there,” he says, “but people who successfully execute are much rarer.”
Disrupting Before Being Disrupted
The same philosophy of continual reinvention can be applied to IBM itself.
At more than a century old, IBM has survived numerous generations of technological change. Syed argues that longevity has depended on the company’s willingness to disrupt its own business before someone else does.
Alongside contemporary investments in hybrid cloud and AI, he points towards longer-horizon areas including quantum computing, materials science and AI safety.
Many of those investments may take years to reach their full commercial potential, but maintaining that long-term innovation pipeline is part of what has allowed IBM to repeatedly reinvent itself.
The other crucial source of insight is proximity to clients.
IBM’s early warning system, Syed says, is not simply a trends report. It is the thousands of engagements taking place with enterprises around the world.
IBM has also sought to demonstrate its technologies internally, an approach Syed refers to as becoming “client zero”.
He says IBM has generated $4.5 billion in run-rate savings over the previous three years and deployed more than 150 generative AI agents into production at scale within 18 months.
“These are not marketing claims,” he says. “These are proof points that show that what we can tell our clients, we’ve done to ourselves as well.”
PepsiCo: Turning Transformation into Intelligent Operations
For a global business operating across multiple markets, supply chains and consumer categories, consistency can become a competitive advantage. PepsiCo’s transformation with IBM has focused on creating exactly that: a more connected operating model capable of supporting faster decisions, smarter planning and increasingly intelligent operations.
The programme entered a defining phase in 2019, when PepsiCo began modernising a technology estate that had developed market by market and function by function over decades. Working with IBM and SAP, the company set out to simplify that environment, standardise processes and create a more consistent data foundation across the business.
In the UK alone, 13 legacy systems were retired, with a similar number removed in Poland during the programme’s early stages. PepsiCo also brought together around 300 subject-matter experts and 34 global process leaders to help shape the solution, ensuring the transformation was developed with the business rather than imposed upon it.
That involvement has been critical to creating a platform that is not only more efficient, but increasingly capable of supporting AI at scale.
For PepsiCo, the value of that consistency can be seen across areas such as demand forecasting, procurement and supply-chain planning. A more harmonised environment enables different markets to work from common data structures, process definitions and performance measures, making it easier to respond to changing demand and manage operations across a complex global organisation.
“When you have that consistency, AI stops becoming a pilot and becomes an operating capability.”
The impact is particularly significant within the supply chain.
“When you have that consistency, AI stops becoming a pilot and becomes an operating capability.”
As AI capabilities continue to mature, PepsiCo could increasingly move from using intelligent systems to provide recommendations towards allowing them to support or carry out operational actions within defined controls.
That could include rebalancing inventory, triggering replenishment or helping adjust trade promotions in response to changing market conditions.
The longer-term opportunity is a supply chain that becomes progressively more responsive and self-optimising, continuously sensing demand, modelling different scenarios and adapting plans as circumstances change.
Yet the transformation is not simply about systems.
PepsiCo’s decision to involve subject-matter experts and super users from the outset has also helped create long-term knowledge and ownership within the organisation. As AI becomes more embedded in everyday operations, that human element will become increasingly important.
Employees need to understand and trust the information behind intelligent systems, while also adapting how they make decisions and work alongside new technologies.
“Once people trust the data and they see AI making their work better rather than threatening it, adoption becomes a pull rather than a push.”
That combination of technology, data and workforce readiness could also support PepsiCo in areas beyond traditional operations.
The same foundations have the potential to contribute to sustainability priorities including water, packaging, emissions and regenerative agriculture, where large volumes of operational data can help identify efficiencies and improve decision-making.
For PepsiCo, the significance of the transformation is therefore not simply that it has modernised legacy technology.
It has created a more consistent global foundation from which the business can operate, adapt and increasingly apply AI where it delivers practical value.
As enterprise AI moves from experimentation towards everyday operations, PepsiCo’s experience shows that the competitive advantage may not come from adopting the newest technology first, but from having the right foundations in place to use it effectively.
“Once people trust the data and they see AI making their work better rather than threatening it, adoption becomes a pull rather than a push.”
Becoming Invisible
Looking ahead, Syed’s ambition for the IBM and PepsiCo partnership is somewhat paradoxical. He wants it to become less visible.
Far from suggesting that the relationship should diminish, he believes the most successful technology eventually disappears into the everyday operation of an organisation.
The ultimate goal is for the platform underpinning PepsiCo to become an almost invisible nervous system, continuously sensing, deciding and learning without the organisation having to consciously think about the technology beneath it.
Agentic AI represents one opportunity, with intelligent systems potentially operating across supply chain, procurement and consumer engagement under robust governance.
Sustainability represents another. Syed sees scope for the same data foundation to support areas including water, packaging, emissions and regenerative agriculture, challenges that increasingly become questions of optimisation at enormous scale.
Looking further into the future, hybrid cloud and quantum-accelerated computing could eventually help address highly complex problems in logistics, formulation and packaging materials.
Yet Syed ultimately believes the significance of the PepsiCo story extends beyond either company. It provides a potential blueprint for how a global enterprise can harmonise its data, create an AI-ready foundation and deploy intelligent technology at scale without sacrificing agility.
As businesses everywhere race towards an AI-enabled future, the lesson from PepsiCo’s experience is that the organisations most capable of exploiting tomorrow’s technologies may be those concentrating hardest on getting the fundamentals right today.