NTT Data’s artificial intelligence guru Bill Wilson on separating AI theatrics from real value creation

Put simply, Bill Wilson sees two jobs for AI in most organisations: first, improving personal productivity; second, the more complex task of tackling company-wide operational and strategic problems.

“The simplest way to think about my role is that I help clients understand how to apply the power of AI to their specific problems,” Wilson says. “Sometimes they do not need AI. Sometimes they just need a data solution or something even more straightforward.”

Given that most firms now have access to tools such as Microsoft Copilot, Gemini or Amazon Q, many professionals now have AI that can help them with ‘unstructured knowledge work’, such as drafting reports, summarising meetings, or building presentations.

“A lot of businesses, even small ones, are buying into those solutions because they can see the benefits,” Wilson comments.

The harder challenge, he explains, is not technical but behavioural. That is, getting people to change how they work. “How do you get people to make AI their default tool for certain tasks?” he asks. Changing habits and workflows has proved to be one of the biggest hurdles for most organisations so far.

Once companies move beyond simple productivity gains, the opportunity becomes much larger. This is where sectors such as real estate and infrastructure are starting to pay attention. These industries generate vast amounts of information, from energy use and maintenance logs to tenant feedback and asset performance data, yet much of it remains scattered and unused.

To fix this, Wilson’s team at NTT DATA is developing systems that allow users to “chat with their data”. The idea is simple. Instead of sifting through multiple spreadsheets or dashboards, people can ask open-ended questions and let AI pull, interpret and combine relevant information.

“The AI not only writes the queries for you,” Wilson explains. “It starts to extrapolate, offer options, and form opinions based on both structured and unstructured data.”

It is important to note that this is not about reading one spreadsheet; rather it is about collapsing fragmented datasets and eliminating manual reconciliation, which is time consuming – and by extension, expensive.

“One thing that is different is that we can synthesise unstructured information and structured information,” Wilson adds. “We are looking at customer feedback and sentiment and fusing that with sales. The AI combines that and makes recommendations.”

In practice, this means reducing the time it takes to produce evidence-backed, robust analysis from a question of days or weeks to mere seconds.

Wilson explains that this same principle applies at the level of entire portfolios or operating businesses. Without consistent and connected ways of processing the vast amount of information generated in real estate, he warns, insight is little more than guesswork. And given that data increasingly underpins the decisions of investors and lenders, he describes this as a critical weakness.

“It is important that you understand how those data structures work,” Wilson says. “When you start to layer on the operational constraints around properties, you need to bring the operational data together with the proper descriptions of those assets. Otherwise, you cannot know how those assets are functioning.”

He gives a simple example of this problem. “In our Old Street building, we have one gas boiler that covers all the tenants,” Wilson says. “If one tenant wants to make efficiency improvements, you do not want that to be at the expense of the other tenants.”

NTT itself offers a live demonstration of this principle. As one of the world’s largest data centre operators, it must balance sustainability, cost and operational performance at scale. As such, efficiency is not just a virtue but an economic necessity.

“We are very careful about how power is used in our data centres because that comes off our bottom line,” he says. This means close scrutiny of every detail of performance. “There is a huge amount of work that goes into optimisation. Not just how the racks of equipment are operating, but how effective the facility is achieving ventilation requirements.”

Wilson points to NTT’s London Data Centre 2 as an example. “The cooling is entirely water-based. No compressor units,” he explains. Another of NTT’s data centres in Munich recycles energy to power a district heating system, though Wilson caveats that district heating is  harder to achieve in the UK where building ownership is more fragmented.

The importance of efficiency also applies to national energy systems, where limited grid capacity constrains both electrification and on-site generation. To show how this affects individuals, he uses his own household as an example. “I have 5.4 kilowatts of solar panels, and I can only put 3.4 kilowatts into the grid,” he says. Local restrictions, he notes, force counterintuitive outcomes. “I end up throwing electricity away in the summer.”

This leads him to argue for smarter use of battery storage and vehicle-to-grid technology as ways to enhance capacity and reduce waste. “We have a lot of spare battery capacity sitting on people’s drives,” he explains. “Modern battery cars are capable of putting energy back into the grid – but don’t.” He likens it to much of public policy, where practical solutions are often blocked by regulation or inertia, leaving an obvious question: why not use that capability?

Another use case for AI in property, Wilson argues, is site selection. He describes a project in which his company helped identify optimal locations for electric vehicle charging stations in central London. This involved integrating many different forms of data – from traffic flows to existing charging infrastructure – and using AI to uncover patterns that would otherwise have been invisible.

“We layered different types of geographical information and created an optimisation algorithm to recommend locations,” Wilson explains. Where previously such an exercise would have required extensive manual analysis and produced limited certainty, the deliberate application of AI achieved more accurate results in a fraction of the time.

Despite the market hype, however, Wilson’s fundamental belief in the underlying technology keeps him grounded.

“If you look at the financial aspects of the AI industry, there is a bubble,” he says. “The question is how big and what is the impact when it bursts.”

“I do not think it is so enormous that we will see a dot-com-style bust,” he adds. One argument being that real underlying value is already visible as revenues pour in, in a way that simply didn’t happen during the dot-com bubble. “We are seeing tangible productivity improvements,” Wilson notes.

Wilson goes on to use the example of the tangible, real world results that have come about from NTT’s work, like for example the use of applied AI in healthcare. “We run a service for the Royal Marsden Hospital that helps advance AI solutions for finding early signs of cancer.” AI is clearly creating value, even if some companies’ valuations seem increasingly detached from it.

Looking ahead, Wilson says the next phase will centre on ‘agentic systems’ – networks of specialised AIs that work together to solve problems.

“These days we have different individual AIs that have specific jobs,” he says. “We can achieve huge amounts of intelligence by using those agents in a team. We are on the cusp of putting agentic solutions into organisations. It is the next stage on from personal productivity.”

As the podcast draws to a close, Wilson returns to first principles, explaining that understanding people is just as important as understanding code. “I spend a lot of time trying to understand customer problems,” he says. “Technology is a secondary question.” As such, progress depends less on technology than on people – on how teams adopt it, what incentives they have, and how clearly the purpose is communicated.

“Sometimes we have made things more complicated than we need to,” Wilson notes. “Sometimes the simplest solution is the best.”

All in all, Wilson provides clear guidance. For leaders in real estate and beyond, the lesson is to start with practical use cases that remove friction, invest in trustworthy data foundations, and focus on real business outcomes. Or, as Wilson puts it, “We always start with the problem statement. Then we decide which technology applies. And often, that is the difference between theatre and delivery.”