The pre-conference dinner at the DIGIT Financial Services Technology Summit brought together senior leaders from across the industry for an evening of candid conversation. No presentations, no agendas. Just an honest discussion about where AI is really taking financial services, and what it will take to get there well.
Seven themes emerged throughout the evening. Here’s a summary of where the conversation landed.
AI adoption is broad, but governance is lagging behind
AI is already embedded across much of the industry. The tools are out there, people are using them, and in some organisations usage rates are remarkably high. But broad adoption does not mean consistent adoption. Even within the same team, there can be significant variation in how people are using AI, how much they trust it, and how deeply it has changed the way they work. Some people are using it to fundamentally rethink how they approach tasks; others are barely touching it. That gap creates its own problems, from inconsistent output quality to growing divides between those who are building AI fluency and those who are falling behind.
What has not kept pace with any of this is governance. Frameworks, policies and oversight structures are being built reactively, after the fact, rather than ahead of the curve.
This matters in any industry, but it matters more in a regulated one. The risk of employees using unsanctioned tools is not theoretical; it is happening now. And in an environment where compliance and data privacy are non-negotiable, the gap between usage and governance is a genuine vulnerability.
The next stage of adoption is less about access and more about intentionality: sharing what works, building consistent practices, and making sure the right guardrails are in place before scale creates problems that are hard to unwind.
Legacy architecture isn't just a tech problem; it's an AI problem
Legacy technology has been a persistent challenge for financial services for years. But the AI era has sharpened the stakes. Older systems are increasingly difficult to understand, with the institutional knowledge needed to navigate them retiring alongside the people who built them. This is a problem that compounds over time.
A significant part of this challenge is structural. Most large financial services organisations have grown through acquisition, and the legacy of that is a patchwork of platforms, data models and systems that were never designed to work together. The result is duplication, inconsistency and fragmentation that is difficult and expensive to untangle. It’s also, increasingly, an AI problem as much as a technology one.
The fragmented data landscape that legacy architecture creates is one of the biggest blockers to realising AI value. Siloed systems, inconsistent data, platforms that don’t communicate with each other: these aren’t just technical inconveniences. They limit what AI can actually do, and they introduce real risk when data that was never intended to be surfaced ends up in places it should not be.
The opportunity, though, is real. AI has genuine potential to help organisations understand and modernise their existing estates. The question is whether firms can move quickly enough while the window to capture that knowledge is still open.
ROI needs to be real, not assumed
One of the most consistent threads through the evening was a collective impatience with AI investment that can’t demonstrate its value. The industry has a long history of business cases that promise returns and reviews that never happen, which is a pattern that can’t continue as AI spend scales.
The firms that will come out ahead are those that move from enthusiasm to accountability. That means defining what success looks like before a project starts, measuring it honestly when it ends, and being willing to change course when something isn’t working. AI for its own sake is not a strategy. The technology is only as valuable as the outcomes it genuinely delivers.
The customer experience gap is wider than most firms think
There’s a meaningful gap between what financial services firms believe they are delivering to customers and what customers, particularly those who are less digitally confident or more vulnerable, actually experience. AI has the potential to close that gap significantly. But it also has the potential to widen it if the focus stays on internal efficiency at the expense of genuine customer outcomes.
The stakes are high. Trust in AI-driven services hasn’t yet been seriously tested in this sector. When it is, confidence can fall quickly and takes a long time to rebuild. The organisations that invest now in getting the customer experience right, including clear human accountability at the moments that matter most, will be in a much stronger position when that moment comes.
The future team is small groups of humans working alongside many agents
The question of what teams will look like in five years came up repeatedly, and the emerging picture felt increasingly plausible to the group. Small numbers of people working alongside multiple AI agents, with human effort concentrated on orchestration, judgement and quality rather than execution.
This is a vision of rebalancing rather than replacement. But it raises a genuinely open question about what kinds of people will thrive in that model. Will AI push individuals towards being more generalist, able to contribute across a wider range of tasks than their specialism would previously have allowed? Or will specialists become more important, specifically because someone needs to validate and take responsibility for what AI produces in complex domains? The honest response is probably that both will be true in different contexts. What is clear is that the skills mix organisations need is shifting, and those that start planning for it now will be better placed.
The ability to work effectively with AI, to direct it, interrogate it and take responsibility for its outputs, is already becoming a differentiator. Organisations that treat this as a core capability and invest in building it, rather than assuming it will emerge naturally, will be better equipped for what comes next.
AI amplifies culture, for better and for worse
The most resonant idea of the evening was that AI doesn’t transform organisations. It accelerates whatever is already there. Firms that are good at change, that have clear leadership and a culture of genuine accountability, will find AI extends their advantage. Firms that are not will find those weaknesses harder to paper over.
This reframes the AI readiness question in a useful way. It’s not really about technology. It’s about organisational health, and whether a firm has the leadership and change capability to use the technology well.
Ethics and accountability need a proper home in the org chart
As AI becomes more embedded in how financial services firms operate, the question of who is actually responsible for its ethical use becomes more pressing. At the moment, the honest answer in most organisations is that nobody owns it clearly, which will need to change.
Part of this is about governance structures, but part of it is also about scope. The ethical questions around AI are broader than most firms are currently accounting for. There’s a real concern about the environmental impact of AI, particularly the energy and water consumption that large-scale model usage requires. For an industry that has invested heavily in ESG commitments, the tension between AI ambition and environmental responsibility isn’t a distant problem, and deserves a place in the governance conversation rather than being treated as someone else's concern.
Where the evening landed
The discussion was honest, wide-ranging and at times genuinely contested. There was scepticism of hype, a strong appetite for accountability, and real thoughtfulness about what the AI transition means for people, not just for organisations.
The question that remains open in the industry is where AI ultimately goes; whether it settles into the background as a capable but unremarkable utility, or proves to be something far more transformative than that. Both views have merit, and the debate is ongoing. What is clear is that internal adoption and customer-facing AI are not the same problem. They carry different risks, and will not unfold at the same pace.
The firms that will thrive are those that treat AI as a genuine change management challenge, not a technology one. The human premium does not disappear. It concentrates around trust, judgement and accountability.
FAQs
Meet the author
Simon Hull is Head of Financial Services at CreateFuture. With over 20 years in banking and wealth, including UBS, Barclays, BlackRock and Deutsche Bank, he helps firms turn AI strategy into practical, accountable change.
Industry Insights
Explore the latest thinking from our industry and tech experts.
AI in UK wealth and pensions: 5 themes reshaping customer experience
Bridging the Gap: Digital’s role in revolutionising pensions
Top Five CX Themes in Financial Services: Key takeaways from the CX in FS conference 2025