Speech
Good morning. It is a great pleasure to be here today at the DAMA Data & AI Conference. This event is about sharing practical experience of building trusted, responsible and effective data and AI capabilities, so I thought it would be a good opportunity to share some of the work we are doing on this at the Bank of England.
In recent years, my colleagues at the Bank have spoken publicly about the implications of AI on the wider economyfootnote [1] footnote [2]. Today, I want to talk about this from a slightly different perspective: what data and analytics transformation means for the Bank of England itself, as an organisation.
But first, let me introduce myself: my name is Ben Martin and I am the Bank’s Chief Data Officer. I took up this role just under a year ago, following a career across the Bank’s economics, markets and supervision areas. This experience gave me a deep understanding of the Bank’s data needs. Taking on the CDO role has allowed me to bring those perspectives together at an enterprise level. It has been encouraging to see how much progress we have made – and continue to make – in improving our data, analytics and AI capabilities. Even though we are making good progress, this is by no means job done. The pace of change in this field is rapid, and our ambitions need to keep pace with it. So today, I want to reflect on the journey we have been on, where we are now, and some of the opportunities and challenges that lie ahead.
Our mission at the Bank of England is to promote the good of the people of the United Kingdom, by maintaining monetary and financial stability. Across our offices in London and Leeds, our people are working to deliver that mission. Central to this is our ability to make the best possible decisions using the most effective data and analytics to achieve our public policy objectives. Data and analytics are integral to our work. They support our understanding of the economy, our supervision of the financial firms within it, and the operations through which we fulfil our responsibilities.
A couple of years ago, we published our Data & Analytics Strategyfootnote [3]. Our overarching aim is to make the best use of data and analytics to fulfil the Bank’s mission. Our ambitions are summarised by three missions.
- First, making it easy for people in the Bank to work with and analyse data.
- Second, bridging data gaps to increase the value of what we collect and share.
- Third, enabling safe and effective innovation, including artificial intelligence.
Those missions rest on two foundations: a modern enterprise data platform on the cloud; and investment in business change, skills and culture to embed this.
That combination is important. We are not simply replacing systems. We are seeking to change how people find and use data, how they collaborate on it with each other, and how they develop new analytical tools and techniques to gain the best insights to make good decisions.
Today I want to focus on how we are enabling, equipping and empowering our organisation to deliver this strategy. Let me give a quick overview before we dive into each in more detail later on.
First, what do I mean by enablement? This covers the work we are doing to build solid foundations by providing transformed systems and high-quality data for our colleagues.
Second, we are equipping our colleagues with new capabilities including AI, with effective guardrails and associated training, so they can safely achieve new insights.
Last, but by no means least, I will cover empowerment – our approach to supporting colleagues to make the most of the investment we have made and to have the confidence to explore these new tools.
I am a big fan of the television programme Grand Designs, watching people use exciting, modern solutions in an innovative way to build something truly special – not just special on paper or to look at, but special for the people who live there. I sometimes think about our strategy in the same way. We are using modern techniques and the latest thinking on how to use data, but not as an end in itself or to wow anyone. Instead, the point is to deliver the Bank’s mission effectively and to make sure our colleagues can flourish. We are enabling colleagues by putting the foundations in place, equipping them with new capabilities to continue building on those foundations and empowering them to make the most of those new capabilities. Thankfully, the one thing we don’t have is Kevin McCloud turning up every few months to see how we’re getting on!
I will say a bit more about the work we are doing on these three Es – enable, equip and empower – at the Bank of England.
Enable – by building solid foundations
As I have said, we are building solid foundations, so we can enable our colleagues by providing effective systems and high-quality data as the bedrock of good analysis. At the Bank, we are part way through a multi-year investment programme to build solid data foundations. This includes investing in new systems and technology, but also continuing to invest in the ‘bread and butter’ of good data management: quality, governance and trust. AI may be grabbing the headlines, but AI runs on data. This audience knows well enough that without high-quality, well-governed and trusted data, even the most sophisticated AI will struggle to deliver. So I will share some reflections on how we are bringing together the new systems and the good old basics to ensure solid foundations.
New systems
At the foundation of our Data & Analytics Strategy is a new Enterprise Data Platform, or EDP. This is our central, cloud-based data and analytics ecosystem. It will bring together a wide range of data needed across the Bank’s responsibilities, including macroeconomic and financial market data we bring in from partners, and the regulatory and statistical returns we collect directly from the financial sector. But the EDP is about much more than storing data. It provides a structured and governed way of bringing data into the Bank, and transforming it into trusted, reusable data that can be readily used for analysis. Alongside this, the EDP provides modern analytical tools and capabilities, including access to approved AI models, helping colleagues to work with data more effectively and strengthening analysis and decision-making across the Bank.
Earlier this year, we deployed the core of the platform. We are now migrating data services and data systems onto that foundation in stages. Those new data systems are being completely transformed from the legacy systems we are retiring. Alongside this technology, we are building the operating model needed to support it. Together, the platform and its operating model are intended to help us move beyond data silos, establish a shared foundation for future data initiatives, and make greater use of cloud-native capabilities to develop new products and services.
The word ‘enterprise’ really matters here. The Bank of England brings together a number of different responsibilities within one institution, in part because of the connections and synergies between them. Our approach to data should reflect that same principle. Data collected or created for one purpose can often have value elsewhere, and insights can be richer when we are able to connect information from across the organisation. So rather than recreating a collection of separate systems, we are building shared data and analytical capabilities that can be used right across the organisation to support delivery of the Bank’s mission.
Good basics
Now, there is no point in building new systems but immediately filling them with poorly managed data. I know that data management is familiar territory for this audience. DAMA gives us a useful way of thinking about it: not as a standalone control, but as the coordinating mechanism that connects ownership, quality, metadata, architecture and use. For us, the real test is whether those disciplines show up in the everyday experience of colleagues working with data. That is why, since launching our Data & Analytics Strategy, we have focused on putting those practical data management foundations in place.
We have an organisation-wide data management policy and an ownership and stewardship framework, alongside data management learning within our wider data fluency programme.
We have numerous Data Owners and Stewards across all areas of the Bank who support over 2,000 registered data assets. These arrangements combine accountability for data with the day-to-day work needed to manage it.
These are important foundations helping us improve consistency, transparency and trust across an increasingly complex data estate. But the existence of a policy, a register, or an assigned owner, is far from being the end of the task. Our objective is that data can be found, accessed, trusted, understood and used, where data management acts as an enabler, not a barrier.
That is the practical test of data management. Does it help someone coming with fresh eyes establish what a dataset contains? Or understand whether it is appropriate for their purpose? Do they know who can answer their questions about it? And can they use it with confidence?
Like other long-established institutions, we face particular challenges in modernising our data estate. The Bank has a long history and with that comes years of accumulated data, systems and processes. Recent independent reviews by the Independent Evaluation Officefootnote [4] and Dr Ben Bernankefootnote [5] have reinforced the importance of tackling these challenges. It is clear we cannot simply build something new and expect all that complexity to disappear. Our Enterprise Data Platform is a good example – as we move datasets across, we are ensuring we look carefully at what we are bringing with us, to improve our data quality and embed those enduring data management basics from the outset. To stick with the theme of houses, this is a bit like moving house. If you live somewhere for a long time, you inevitably accumulate things along the way, which get shoved in drawers or the loft. Moving gives you the opportunity to decide what you really need, tidy things up and organise them better, rather than simply packing everything into boxes and recreating the same clutter in your new home.
Equip – by providing new capabilities
Next, we want to equip our colleagues to safely achieve new insights by using new capabilities including AI, with effective guardrails, and by providing the right training.
New capabilities
Much of the focus on data tooling over the past couple of years has been around the opportunities from new AI capabilities. We have been open about our approach to using AI at the Bank of England, which we outlined in our AI strategy that was published last yearfootnote [6]. To support a consistent and coherent approach to AI across the Bank, our AI strategy rests on a set of principles that set the tone for how we approach AI. Although frontier AI models and their capabilities have advanced rapidly over the past year, these principles are enduring. The principles are embodied in the ‘TRUSTED’ framework, which is for our AI solutions to be:
- Targeted on tightly defined use cases that have a clear business purpose.
- Reliable with high performance standards and grounded in high-quality data.
- Understood in the context of their intended purpose.
- Secure so we protect ourselves from risks such as unauthorised access.
- Tested thoroughly for their technology standards and impact on behaviour.
- Ethical including being beneficial and scientifically rigorous, fair and inclusive, transparent and secure, and compliant and accountable.
- Durable so they remain effective despite changing conditions and usage patterns.
To bring this to life, I thought I would share some examples of how we have implemented this framework in practice. Our approach combines broad access to general purpose tools across the Bank with a more tailored applications for specific business needs.
A significant milestone has been the organisation-wide rollout of Microsoft 365 Copilot. Nearly all colleagues at the Bank use it each week. Some colleagues also use GitHub Copilot to support code development. As colleagues have become more confident, this has extended to the creation of many custom Copilot agents over the past year, typically to improve personal productivity in day-to-day tasks or to help navigate internal services. Many began as experiments, but some are now being used more widely across the organisation. For example, our ‘AskHR’ agent provides answers to questions on leave, pay and our people policies. Another agent, called ‘AskTech’, helps with technology-related questions and can direct colleagues to further support where needed. As you can see, our approach to naming them has been slightly less experimental!
Alongside these general-purpose tools, we are also developing bespoke AI applications for important business processes.
One example is our Agency Retrieval and Insights App, known as ARIA. The Bank’s regional Agents, based in 12 agencies across the nations and regions in the UK, maintain close relationships with businesses of all sizes, providing the Monetary Policy Committee with real-time insight into economic conditions and perspectives from across the country. Those relationships and the judgements of our Agents remain at the heart of that work, but ARIA complements this by helping them make sense of this wealth of intelligence they gather, using capabilities such as theme detection and quantitative analysis to identify patterns and draw out insights. It is a good example of technology augmenting human expertise and strengthening one part of the evidence base that supports the MPC in making its policy decisions.
Another example is called GRACE, which provides retrieval-augmented AI to enable colleagues to draw deep insights from large and trusted internal knowledge repositories. But what I particularly like is that it does this through an email interface, a tool colleagues already know how to use every day. They can send off a query, get on with something else and come back to the response when it arrives. The interaction is also captured naturally in the email chain. It is a simple example of designing AI around people and the way they work, rather than the other way round. These applications address different needs. What connects them is the effort to make technology useful within the work people already do.
Effective training
The pace of technology change can be dizzying, and with new capabilities emerging all the time, it can be difficult to know where to start. That is why we need to equip colleagues not just with the access to these tools, but to also equip them with the confidence and skills to use them safely and effectively. We want to help them feel confident navigating that change and so we have maintained a focus on training which is one of the two foundations of our Data & Analytics Strategy. It focuses on business change, skills and culture, recognising that successful adoption requires practical support, as colleagues move to new services and learning that goes beyond narrow technical training.
To put this in practice, we have a dedicated data and analytics learning function. To date, I am proud to say it has delivered more than 250 courses to many colleagues across our data science and analytical toolset.
Our aim is not simply to provide more training. It is to make learning more relevant to the work people need to do. On Copilot, we began with introductory learning to establish a common baseline for everyone. Many colleagues then attended application-based deep dives. We have subsequently introduced clinics offering more personalised support for local and role-specific use cases.
That progression from training course to back at the desk matters. General awareness can help people get started. Practical support helps them move from experimenting with a tool, to using it effectively and achieving results.
But formal learning is only part of the picture. We have a thriving Data Community, with almost 2,000 members. It actively supports collaboration, sharing and peer-to-peer learning. During our ‘Data & Analytics Fortnight’ in June, over 1,000 colleagues turned out to one of our events, which included a wide range of internal and external presenters sharing their expertise. Beyond this, the community also organises regular hackathons which have turned out some brilliant solutions.
Bringing people together is incredibly important but the value of a community is not only in the events it organises. It is in the opportunities it creates as a place for people to ask questions, share experience, and learn from colleagues who have faced and overcome similar challenges.
Beyond catalysing an organic approach to building capacity, we are also developing more structured pathways for embedding data skills professionally.
We launched the Data Profession last November, bringing together colleagues from across the Bank whose work involves data into a single professional community. It covers disciplines including data science, data analytics and data governance, helping us take a more joined-up approach to the data skills we need, how we recruit and develop people, and the career pathways and support available to them.
Building our capabilities also means broadening access to opportunities. I am pleased to say that our recently established partnership with Women in Data helps this through a range of initiatives supporting our commitment to greater representation and inclusivity in the workplace.
Like our data tools and technology, our learning approach must also keep evolving. We are working to understand shifting skills requirements across different time horizons, using a role-based approach to assessing skills profiles to help us to map the work different colleagues need to do, against the skills required to do it effectively, and the most appropriate ways to support them in addressing any gaps.
So, we are equipping colleagues to safely achieve new insights by using new capabilities including AI, with effective guardrails, and by providing the right training.
Empower – by supporting colleagues to innovate
Investing in equipping and enabling our people is important, but to unlock its full potential, we need to empower colleagues to try innovative new approaches in their roles. This is not just about doing what we already do more efficiently. It is also about opening up new possibilities, helping colleagues work differently and undertake types of analysis that simply were not possible before. Ultimately, it is about giving people the tools and confidence to evolve with the world around us.
Support
Like many organisations, we want to reduce the costs to trying new innovative approaches and to avoid the risks of this happening in siloes. To do so, we are bringing together a dedicated team of experts from both our data and technology functions to act as an innovation hub for the Bank as a whole. The aim is to support colleagues to understand the range of tools available to facilitate innovative work and to support them to bring insights from across different teams. The idea is not that we try to do the innovative analytical work centrally, but instead to be a centre of excellence and smooth the way for the innovative work that the various business functions of the Bank want to achieve. This includes a forum where local business and central support functions’ representatives can meet, share developments, and to prioritise work to achieve the biggest benefits.
Culture
Finally, we need a culture that gives colleagues the confidence to try something new. Sometimes trying something new feels daunting. We want colleagues to feel that innovation is part of their role – that they are empowered to question how something is done today, explore whether there might be a better way, and have a safe environment in which to test new approaches, which can be supported by the innovation hub. It gives colleagues somewhere to experiment with the right support and safeguards around them, reducing some of the barriers to trying something new. Of course, the design of this needs to be appropriate for the evolving culture of the Bank, just like a well-designed house needs to work in the place where it’s built and for the people who live there.
Senior leaders have an important role here too. We need to visibly take the lead in encouraging people across the organisation to innovate, make space for them to do it, and celebrate the examples that demonstrate what is possible. Over time, that is how we make innovation feel more like a natural part of how we work, and why we are launching in the coming months a new leadership programme focused on leading in an increasingly digital, data and AI enabled world.
Conclusion
Let me conclude by returning to the main theme of my remarks today. Data and analytics transformation is not primarily about technology for its own sake. It is about making sure the Bank of England can continue to fulfil its mission effectively in a world where the volume of data, the sophistication of analytical tools, and the pace of technological change are all increasing rapidly.
For us, that means doing three things together. First, we need to enable colleagues by putting the right foundations in place: modern platforms, well-managed data, clear ownership and the enduring disciplines of data management. Those foundations matter because high-quality analysis, effective decision-making and responsible AI all depend on trusted and accessible data.
Second, we need to equip people with new capabilities and the confidence to use them safely. That includes making general-purpose tools such as Copilot available at scale, developing more tailored applications where there are specific business needs, and applying clear guardrails through our AI framework. But it also means investing in training, communities and professional pathways, so that people can use these tools well and understand their limits.
Third, we need to empower colleagues to innovate. The opportunities from data, analytics and AI will not be realised by a central team alone. They will come from colleagues across the organisation identifying better ways to do their work, experimenting responsibly, sharing experience, and building on each other’s successes. Our role centrally is to reduce the friction, provide expertise and make it easier for good ideas to move from the drawing board to practical use.
We cannot stand still and our job is not done. At the Bank, our ambition is to build on the progress we have made and keep moving forward: enabling colleagues with trusted foundations, equipping them with safe and effective capabilities, and empowering them to innovate in support of the public good.
The technology around us will continue to evolve and expectations of what is possible will continue to rise in the era of AI. But keeping pace is not just about technology. It is also about continuing to learn from one another. The work of communities such as DAMA has a vital role to play in bringing people together, sharing practical experience as we are today, and ensuring that even as technology moves quickly, the standards underpinning our data remain strong. Thank you.
And if Kevin McCloud is in the audience, ignore what I said earlier – you're welcome to come to the Bank any time you want!
Acknowledgements
I would like to thank Aimee Newmarch and Miranda Hewkin Smith for their assistance in drafting these remarks. I would also like to thank, Nay Assassa, Karen Bassett, David Bradnum, Jane Cathrall, Benjamin Crampton, Elizabeth Hall, Rachel Heppinstall, Alan Hicks, Rebecca Jackson, Sarah John, Neil Kisserli, Georgios Kyriakopoulos, Leslie Lambert, Clare Lombardelli, Phillippa Loveridge, Pat O'Connor, Arshadur Rahman and Paul Robinson for their input and helpful comments.