Blog

A recent roundtable discussion revealed that financial institutions may be softening their long-standing resistance to core IT renewal – as AI raises the stakes and exposes the limitations of legacy systems.

The debate, hosted by SoftServe and Google Cloud at London’s House of Lords, brought together senior technology and operations leaders from across financial services.

The legacy paradox

The conversation started with a central question: “Is something considered legacy because it’s old, or because it’s a problem?”

Attendees quickly defended their existing systems. One pointed out that legacy systems often provide a competitive advantage – refined over decades to solve real business problems that newer software doesn’t fully understand. Another noted that systems are never really replaced – layers are added on top of layers until no one is quite sure what the original platform does. That uncertainty itself becomes a reason not to touch it.

Arguments against modernizing piled up:

• Customers don’t care about the back end
• Any project that fails the value-to-client test is hard to defend
• Cost is a perpetual worry, especially when modernization shifts costs from Capex to Opex
• Decommissioning is harder than building and harder still to get approved
• There’s a shortage of success stories – without peers pointing to clean migrations that paid off, inertia keeps winning

There was agreement that replacing systems simply because they’re old has no merit. But the cost of doing nothing changes once new technologies enter the picture.

AI as a catalyst

Some attendees acknowledged that AI might now be the external event that changes the calculation. Several said their organizations were under explicit top-down pressure to deploy AI tools. One participant reported that staff were told their jobs would be safe if they used AI to improve productivity – and at risk if they didn’t.

But that pressure hasn’t translated into universal enthusiasm. Banking hasn’t fully bought into AI, one attendee said, because the industry doesn’t yet trust it. Another was more sceptical, pointing to the hype and pricing that assumes perfect technology – but one that still makes mistakes. “Has anyone really achieved AI success at scale,” they asked, “or are we still just looking at impressive pilots?”

However, some concrete wins were cited. One attendee described using AI to process documents related to car-finance mis-selling claims – a paper-heavy problem that AI handled well. Another said their call centre had moved from reviewing a random sample of calls to having AI flag every conversation that needed human attention – delivering better service without reducing headcount.

The discussion showed that while for some, AI was an excuse to modernize, in many cases it still required a business case. The harder question is often whether the underlying systems can support it.

Questions of trust

Trust underpinned much of the conversation – trust in new technology, trust between technologists and the business, and the customer trust that financial firms depend on. That’s why firms have put such a premium on high security standards.

Regulation always shapes risk appetite in financial services. UK regulators were described as still fearful of firms even trying AI – partly because they want proof that decisions are correct, which non-deterministic technology can’t easily provide. Other attendees were more sympathetic, noting that regulators help firms modernize properly by holding them to standards they might otherwise duck. Internal audit teams, another added, often set a higher bar than regulators anyway.

Most agreed that the world is moving faster than corporate governance can manage. That makes the choice of risk frameworks as important as the choice of technology. Can the new system be governed? Can the AI be controlled?

Two paths forward

Younger businesses can sometimes carry a “technology premium” in their valuation – an investor reward for not being weighed down by mainframes. Older firms can’t replicate that, but they can take a different route.

Incremental change is often more sensible than a big-bang replacement – particularly as the people who have maintained mainframes for decades start to retire. Even here, AI itself may help by writing and refactoring older code, including COBOL.

Agents, humans, and bottlenecks

Agentic AI drew cautious assessments. Roles are already collapsing because of AI, one participant said, but the tools haven’t been put into practice yet. Another noted that running an AI agent full-time is currently more expensive than hiring a data analyst.

The longer-term bet is that AI will commoditize itself, and humans will become the differentiator. A more human-like interaction generated by AI may improve customer experience, but trained people will still close the loop. Processes are owned by people, not technology.

Cloud resources emerged as a potential bottleneck – one attendee said their organization wasn’t scaling cloud fast enough to meet AI ambitions. Data domicile was another concern, as banks need to know where data is stored, and cloud providers can’t always guarantee jurisdiction. Solutions like Sovereign Cloud – physical equipment deployed on client premises – are helping address these concerns.

The closing thought

The pace of change is unlike anything seen before. But the industry has an advantage that earlier disruptions didn’t offer – miners in the industrial revolution had no warning that the steam engine was coming. Today’s banks and insurers can see what’s on the way.

The question is what they will choose to do with that foresight. But the discussion showed that many are now giving core modernization more serious consideration – to better deal with the challenges ahead.