AI in Financial Services: Hype or Transformational Reality?

At FinTech LIVE London 2025, industry leaders explored how AI is transforming financial services. 

In the AI in FinTech Forum at FinTech LIVE London 2025, Sam Bridges-Sparkes, Head of BI Analytics and Strategy at Shawbrook Bank, joined Susanne Chishti, Chair and Founder of FinTech Circle, Babatunde Akinkugbe, Investments and Technology Director at Lloyd's Banking Group and Mark Watson, CTO at ComplyAdvantage, to discuss the opportunities and challenges of AI adoption. 

The panel examined the shift from hype to reality, addressed critical questions around risk management and governance, and debated whether we’re witnessing an augmentation of human capability or the beginning of wholesale workforce transformation.

How has the fintech landscape changed with AI adoption?

Sam Bridges-Sparkes: AI is a huge accelerator for fintech’s ability to scale quickly. On the banking side, where we struggle with technical debt and the inability to change, AI can possibly solve that problem. We’re seeing a lot of convergence in this area, and I think it’ll become much more of a partnership model over time, rather than just a SaaS provider arrangement.

I think SaaS is probably in the biggest danger of disruption. Why pay someone to build simple things when my wife, who isn’t technically literate, can knock up a website and a database in a matter of minutes nowadays? 

The out-of-the-box capabilities of frontier companies like OpenAI and Claude are progressing far quicker than anyone can keep up with in that general space.

I'll add to this idea of speed of delivery becoming a key KPI. Something we're starting to track that we haven't done before is delivery time, because the landscape is forever changing and the hype doesn't seem to fade. I'll give you an example: we did something with AI where we took a process from 45 minutes down to three minutes. 

The first response from the crowd was that it still takes quite a long time. Tough crowd, but that's because it wasn't a long project. The days of multi-year technology change programmes are almost forgotten now, because what we build today is likely already legacy in six months' time.

Mark Watson: From a technical perspective, we use a lot of machine learning models – we had about 30 to 40 of them in production. They’re gradually being replaced by large language models. 

The machine learning models would take about six weeks to retrain if something changed. The large language models take about an hour to reprompt. We're seeing significant possibilities.

What's really changed is the advent of large language models. The potential applications are considerable – they can apply them to natural language, to coding to anything really that can be expressed in words. I think we are at the beginning of the cycle, not even in the middle of it. 

The challenge from an executive point of view is figuring out where you are in the cycle and matching speeds to that, not trying to get too far over your skis in terms of implementing things.

Why is generative AI such a game changer for the industry?

Susanne Chishti: The reason why generative AI is such a game changer is that it makes such a big difference to certain ways of working and certain problems we’re trying to solve. Ultimately, it comes down to the business problems we want to solve and the challenges we face.

Just to give you one example: a technology company I know was facing a hostile takeover from a Chinese company. Instead of hiring a lawyer for £10,000-plus to investigate the relevant laws, they used ChatGPT to get all this legal information. There was so much value in being able to investigate and get high-quality know-how from generative AI.

In terms of FinTech Circle, all of the startups we invest in use generative AI, either for marketing purposes – generating blog posts, webinars, even videos – or for coding. You can use it to cut your coding times. But often the best technology developers tell me you get 70% right, but the last 30% you still have to do yourself. You have to do quality assurance.

One good piece of advice I received is you should only ask generative AI questions where you can quality-assure the outcome yourself. 

Otherwise, if you pass on incorrect information, you might embarrass yourself. This comes to the bigger question of how we train young people and graduates who come out of university. We need to train them to understand what is right and what's wrong, so they can use generative AI properly.

Babatunde Akinkugbe: It's a game changer in the sense that if you look back in history – the Bronze Age, the Industrial Age – we are now in a different technology change era, and generative AI has effectively accelerated that.

From a banking perspective, we've had big banks, new banks, challenger banks all trying to come in, and we've withstood that. But this one, if we do not catch up quickly and embrace agents, is a real existential threat. You have agents talking to agents – why will you need a bank in that space? We could be disintermediated, and that will be the core issue.

It’s not just going to be limited to the banking and financial sector. We’re going to be thinking about society at large. We’re now beginning to seriously consider a universal basic income because people will be out of work. 

There will be some categories of jobs that are just not going to be needed anymore, and we do not have enough skills to meet that demand. The transitional period is going to have some lags.

Are we opening up customers to too many potential breaches and risks?

Babatunde Akinkugbe: The short answer is yes. If you think about it, one of the parts that many are not actually considering is the amount of power AI demands and what it's actually doing to the environment. For each prompt, there’s a significant amount of energy consumed.

When you think about bias, it can quickly become ingrained and scaled significantly. So for any positive aspect, there's also an alternative risk. 

We’re talking about quantum cryptography emerging, and the threat there is not just the positives of quantum computing – it’s that it’s going to destroy encryption. Merge that with AI, and you just scale the problem. We need to start preparing for that.

Susanne Chishti: If anybody has used ChatGPT or any AI, has AI ever disagreed with you? Because it normally never happens to me. It always agrees with me, it always compliments me. That's what AI is trained to do. LLMs are trained to complement you: “Yes, it’s a very good question. Do you want to know more?” It helps you to go deeper and deeper on your journey.

If you are aware of that and you handle the question correctly from the beginning, it’s nice because you get this confirmation that you're on the right track and you're asking the right questions. 

But if you have issues, maybe personal issues, or you’re a vulnerable customer in our financial services sector, it can be very dangerous because you might get into a rabbit hole where you get negative confirmation.

We have seen one person commit suicide in the US from using AI as a therapy tool. That’s the biggest use case globally at the moment for generative AI – to be used as therapy, where people prefer talking to the agent instead of talking to people. Getting the wrong advice is very dangerous. It has lots of risks.

From a bank’s point of view, it’s about training – training employees how to use it, what to put in and making sure that you've got your private framework, which is privately secured AI only for your company, which can't be shared and can't be used to train the main model outside of your firm.

AI alongside humans: augmentation or replacement?

Sam Bridges-Sparkes: AI is very much the sixth rung on a ten-rung ladder. Use it to automate the mundane, augment your current roles – very much as a partnership. I think that's one space where we're seeing it working very well.

I'm a skeptical of chatbots. I think there might be a change shortly where your actual customer service improves when you can talk to people. The first thing I try to do with a chatbot is get to a human. 

Even though I know how the thing works, I still can't get to one. It's about freeing up that human capability to do personal interaction, complex thinking and engagement. 

We shouldn't be taking minutes in a meeting anymore. We should be listening and engaging in that meeting because AI is doing that. For me, I think that's where you need to start and follow that curve out.

Susanne Chishti: Some of those agents are still very poor quality. You probably see that in your own inbox when you get emails from salespeople that are written by an agent. I’m getting so many emails that I think they’ve been generated by an agent who's not properly trained. 

It's so obvious that it backfires, because you would never want to work with a company that uses this type of low-quality approach. So it's very important to check the quality before you engage.

Babatunde Akinkugbe: I think it’s here to stay and it’s only going to deepen. There’s a case of a consultancy company recently, where the client is suing them because they produced a report, sent it over and the client checked it online and found it was written by an agent or a chatbot. This is already happening and it’s only going to scale as we go further. It’s a reality.