Leon Butler

In his office at IBM’s UK headquarters in London, Leon Butler is discussing test tubes. It's an unusual metaphor for the recently appointed CEO of IBM UK and Ireland to use when talking about AI, but it cuts to the heart of his vision for the technology.

“My favourite analogy is an opaque test tube with liquid inside – that’s the data the model’s trained on, but you can’t see what’s in there,” he explains. “You add your own data, mix it up, but would you drink it? You need to understand what the model's trained on before mixing in your data.”

This focus on transparency runs through Leon’s strategy since he took the helm in January, following his return to IBM after four-and-a-half years at Oracle. While many enterprises chase headline-grabbing AI implementations, he is steering IBM’s UK and Ireland operations toward measurable productivity gains and practical applications.

“We are a leading AI and hybrid cloud company and AI is very much integral to that,” he says. “We support enterprise clients as well as the public sector, focusing on productivity gains and making them successful through AI and hybrid tools.”

IBM watsonx tackles fragmented data challenges for enterprises

Six months into his tenure, Leon is confronting what he sees as the primary obstacle to effective AI deployment: data quality. This focus has shaped IBM’s watsonx portfolio, which has evolved from its Jeopardy-winning origins into a suite of enterprise tools.

“Data is one of the biggest inhibitors that we see,” Leon explains. “We’re seeing a lot of unstructured data challenges now within AI in particular.”

The watsonx portfolio now encompasses several components addressing different aspects of enterprise AI implementation. Watsonx.AI provides model development and training capabilities, while watsonx.data focuses on structured enterprise information. Watsonx.governance – its platform to speed responsible, transparent and explainable AI – has gained particular traction among regulated industries.

“Watsonx.governance is top right on the quadrants and you’ll see a lot of global systems integrators using it from a governance perspective,” Leon notes.

The company’s recent acquisition of DataStax highlights this data-centric approach, bolstering IBM’s capabilities in retrieval augmented generation (RAG) through vector databases. As Leon emphasises, solving fundamental data challenges unlocks the potential for return on investment.

“Being able to harness information and get that information clean and accurate is definitely one of the bigger challenges. Getting that right does really help return on investment,” he says.

Internal transformation provides template for enterprise adoption

When discussing successful AI implementation, Leon doesn't begin with customer case studies but with IBM’s own transformation. The company applies what it calls a ‘Client Zero’ methodology, serving as its own test case before deploying solutions externally.

This approach has generated substantial internal benefits. “IBM has generated around US$3.5bn from productivity gains, about US$2bn of that through AI and automation,” Leon says.

One of the most successful internal projects has been Ask HR, a natural language interface for employee queries that now handles 11 million interactions annually. The system emerged from analysis of help desk tickets that revealed inefficiencies in IBM's traditional HR processes.

“We saw things like promotion cases or transferring employees taking a huge amount of time,” Leon recalls. “We were able to analyse that through our help desk tickets.”

The resulting solution combines natural language processing with workflow automation to route and resolve queries. “You can literally ask for something in natural language, and that agent will make a decision to either set off an email or go to the HR system at the back end and actually work that route out for you,” he explains.

The metrics from this implementation illustrate the productivity potential. Help desk issues have been reduced by 75%, while the time to measure or process those particular tasks has again been reduced by 75%.

Perhaps more significantly, the project has transformed roles within IBM's HR function. “Now a number of HR professionals are data scientists or conversational specialists as well, using HR terminology to help us continually streamline that process even more,” Leon adds.

These internal experiences have informed IBM’s approach to client engagements across sectors. In financial services, NatWest Group faced challenges processing 20,000 legal documents annually, a task Leon describes as previously manual and time-consuming.

“We’ve worked with our Consulting Expert Labs and watsonx.technology to have AI analyse those documents to make decisions around when to use them and when not to use them with legal terms,” he explains.

In healthcare, IBM partnered with Hospitals of Warwick and Coventry to address the persistent issue of missed appointments using AI assistants. “They had a challenge of missed appointments, and then we had an AI assistant who was able to not only identify those missed appointments but notify when appointments freed up,” Leon says.

The impact has been substantial in a sector where resource optimisation directly affects patient care. “This helped them benefit by about 700 appointments per week, quite substantial when it comes to reducing the NHS waitlist,” he notes.

IBM Granite models offer alternative to resource-intensive LLMs

While much industry attention focuses on increasingly large language models requiring substantial computing resources, IBM has pursued a different path with its open source Granite model family which targets specific business use cases.

“Granite is a family of small, agile, specialised AI large language models. They’re not the 750 billion parameter models out there,” Leon explains. “You can run them on the cloud or on premise with a smaller number of GPUs.”

This approach delivers practical advantages in cost and efficiency. “Our models are 2-3 times more cost-effective and run more efficiently,” he claims.

The transparency of these models addresses growing concerns about data usage and training sources affecting larger, proprietary systems. “The big issue with proprietary 'black box' models is you don't know what they're trained on. We pivoted early towards transparency,” Leon says.

Lessons from field deployments point to targeted approaches

Drawing from IBM's experience implementing AI across industries, Leon identifies several patterns separating successful projects from failed experiments. The central factor, he argues, is focusing on specific use cases with measurable outcomes.

“It’s all about the use case and quantifiable outcomes. I’ve seen huge numbers of pilots that went nowhere,” he observes. “Where we pivoted quickly was ensuring a use case with quantifiable outcomes, then working backwards from there. It sounds simple, but it’s amazing how many companies didn’t do it.”

This emphasis on practical application extends to implementation strategy. Leon advocates starting with targeted deployments rather than enterprise-wide initiatives. “Start small and grow to gain credibility,” he advises. “Make sure the use case works, which earns the right to ask for further budget.”

He also cautions against automating inefficient processes without reconsidering their fundamental value. “When automating processes, ensure they're processes worth automating rather than just replicating bad processes.”

UK economic potential hinges on skills and data sovereignty

Looking beyond individual implementations to the broader UK economic landscape, Leon identifies several requirements for realising the potential of AI technologies at national scale.

Skills development features prominently. “AI growth zones and skills augmentation across the entire UK and Ireland are crucial,” he says. IBM has established targets through its SkillsBuild program “to skill up two million people in AI by 2026 and 30 million by 2030, open to everyone including schools and academia.”

Data sovereignty also requires attention in AI strategy development. Leon emphasises the importance of “giving clients choice between cloud storage or running behind their firewall with complete data control.”

The potential economic impact justifies these investments, according to government projections Leon cites. “The government believes £400bn (US$533bn) could be added to the UK economy by 2030 through AI. We have the technology to achieve this now.”

As Leon looks ahead to the future of AI adoption in the UK, his outlook blends optimism with pragmatism. “I’m passionate about this because we’re just starting. People are playing with the technology, but focusing on specific use cases and implementing end-to-end processes is where you get real productivity gains. That’s where you augment people to perform higher-level tasks. If we can reach that point across the UK and Ireland economy, it would be a genuinely great place to be.”

To read the full article in the magazine, click HERE.

Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.

Discover all our upcoming events and secure your tickets today.

Technology Magazine is a BizClik brand