AI is transforming how financial firms handle everything from customer engagement to internal operations. Yet in a sector where trust, regulatory compliance, and accuracy are essential, the journey from pilot to production is challenging.
The rapid deployment of AI at scale raises a crucial question for businesses: can innovation keep pace with regulation?
Rob Paisley, strategic industry director – global team lead at SS&C Blue Prism, believes success starts with focusing on high-impact use cases.
In production today, “Most of the use cases around our use of AI are either extracting data from an unstructured source, like a document that you’ve never seen before, or aiming it at a large data set and asking questions of that data.”
These foundational applications – document processing and big‑data querying – are practical, proven and scalable.
From experimentation to real impact
AI is already helping financial services firms move faster and serve customers better. Smarter chatbots, improved self-service journeys and real-time personalisation are starting to deliver tangible benefits across the sector. But while many firms have launched AI pilots, few have moved beyond isolated use cases.
“You see a lot in the press around insurers doing AI projects, but they always seem to be single-point solutions,” says Dan Huddart, chief technology officer at Avantia Group. “We thought very carefully about how we redesign the company to embrace AI and get the benefits of it at scale and consistently across the business.”
Avantia’s approach is underpinned by a simple principle: AI should simplify complexity, not compound it.
“Our ambition with AI is to try and make agents’ roles simpler so they can focus their time on the call, on the customer,” says Huddart. “AI picks up 100% of what was said, processes it, cross-references it against the policy, and gives a very quick indication of what we are covering.”
Building trust through human oversight
While the efficiencies are clear, so are the risks. In a heavily regulated sector, it’s not enough to be fast – systems must also be fair, precise and explainable.
“Our current operational model is that we would always keep a human in the loop for any decision that impacts a customer,” says Huddart. “Our human staff are exceptionally good at empathy, relationship building, confidence – but also judgment.”
Keeping that human layer is also key to building confidence internally. “A lot of people get over their skis with AI,” Paisley warns. “They fall in love with the hype without necessarily knowing if they’re ready or prepared.”
Data debt drags on progress
Many firms underestimate the barriers to scale. Chief among them? Data quality. “Usually that data is not at all in one place. It’s not clean. It’s never as good as you think it is,” says Paisley. “You start talking about containers and multiple releases per day and their IT teams get scared.”
Legacy systems pose a major roadblock – particularly in insurance, where core systems of record aren’t designed for speed or agility.
“Most insurance companies are built on two big systems of record – a policy system and a claims system,” Huddart notes. “They’re designed to be accurate. And they should be. But they’re not designed to be flexible.”
To work around this, Avantia has created a separate platform designed for experimentation and speed. “We move all of our business decisions up into our own separate platform that ingests data at speed, uses machine learning and generative AI,” he says.
AI needs authority and skills
Technology alone won’t deliver transformation – embedding AI across a business requires rethinking talent, structure and leadership.
“The next biggest challenge is this concept of design authority,” says Paisley. “Making sure that when you put a person in charge of designing an AI system they actually have the authority to make changes throughout the business.”
“We deliberately hire about 50% of new entrants from outside the insurance industry,” says Huddart. “We train them in the fundamentals of AI and prompt engineering. And we embed our data scientists into departments like pricing and operations – not in some ivory tower.”
Paisley sees a similar benefit in federated models. “We train up business analysts and embed people from the centre of excellence in every single department,” he says. “A lot of times, the people who find the best use cases are the ones who’ve been doing it manually for their entire career.”
Business-wide transformation
For both leaders, AI isn’t just a technical shift – it’s a strategic inflexion point. “My hope is that it will re-architect the operating model for insurance,” says Huddart. “Insurance hasn’t really ever had the same chance as banking or payments. Hopefully, AI may well be the straw that triggers that.”
Paisley agrees that time will be the next frontier. “We’re going to move away from this nine-to-five economy to this always-on economy,” he predicts. “And that’s going to put a lot of pressure on institutions to be able to turn around things very quickly.”
AI may be the enabler – but trust, clarity and structure are what will turn promise into progress.
The winners in this space won’t be those that move fastest, but those who scale safely, build on solid data foundations, and empower people at every stage of the process.
AI is transforming how financial firms handle everything from customer engagement to internal operations. Yet in a sector where trust, regulatory compliance, and accuracy are essential, the journey from pilot to production is challenging.
The rapid deployment of AI at scale raises a crucial question for businesses: can innovation keep pace with regulation?
Rob Paisley, strategic industry director - global team lead at SS&C Blue Prism, believes success starts with focusing on high-impact use cases.
In production today, “Most of the use cases around our use of AI are either extracting data from an unstructured source, like a document that you've never seen before, or aiming it at a large data set and asking questions of that data.”
