‘There’s little room for caution’: Banks and fintechs gear up for AI revolution

The financial services sector is leaning into the hype around artificial intelligence (AI), as banking and fintech executives hail its potential to boost productivity and reduce costs.

The technology is booming across global financial services. Banks’ spending on generative AI alone is estimated to hit $85bn (£64.3bn) globally in 2030, from $6bn (£4.5bn) in 2024, according to Juniper Research.

McKinsey forecasts it could add up to $340bn (£257bn) in value to the global banking industry each year, or roughly 4.7 per cent of its total revenues.

While still most commonly used for marketing and customer service, an increasing number of financial firms are testing new use cases for AI as the technology advances.

“By analysing vast amounts of data in real time, fintech companies are able to offer hyper-targeted financial products, automate complex processes and anticipate consumer needs with unprecedented accuracy,” Janine Hirt, chief executive of trade body Innovate Finance, told City A.M.

Fintech start-ups are moving quickly to leverage emerging areas of AI in their challenge to bigger incumbents. Hirt said the sector has “revolutionised areas such as fraud detection, credit scoring and customer support through advanced algorithms and machine learning”.

One earlier adopter of ChatGPT was Klarna. The Swedish bank – and UK’s biggest buy-now pay-later (BNPL) provider – said last week that it was planning to nearly halve its roughly 3,800-strong workforce as it hailed the role of AI-powered “efficiencies” in improving its profitability.

Klarna claims nearly 90 per cent of its employees regularly use AI for tasks like drafting legal contracts. In February, the firm said its AI assistant, powered by San Francisco start-up OpenAI, was handling a workload equivalent to 700 full-time customer service agents.

“The people who really thrive with a new technology are the ‘tinkerers’. Those people who, at all levels, start playing, learning and working with new technology quickly,” said Yaron Shaer, chief technology officer at Klarna.

“Very early on, we saw that Gen AI adoption must be the responsibility of the whole organisation, not a single ‘AI transformation’ team.”

He added that providing the whole company controlled access to AI tools helped Klarna “take advantage of the good parts while also understanding where and how to build the right controls and guardrails to avoid the shortcomings”.

While more established banks have been using chatbots in customer service for years, their role has grown significantly as AI technology develops.

Natwest has said its AI chatbot, named Cora, handled 10.8m queries last year, up from around 5m in 2019. City A.M. understands Cora has already handled more than 5.9m customer interactions this year.

The bank is now rolling out Cora+ with IBM, which promises to deliver “a more seamless ‘human’ experience” using the latest AI innovations.

Elsewhere, relationship managers in Natwest’s private bank are said to be saving an average of 15 minutes per customer call through “summarisation tools”.

‘Either you onboard AI or get left behind’

While its potential to optimise products and services is clear, the rise of artificial intelligence has also prompted concerns over potential risks to both consumers and employees.

The large language models which underpin most generative AI tools have been known to “hallucinate” and create false information, while some have warned against training chatbots on sensitive customer data.

Other uses for AI in banking and payments include checking transaction data to catch suspicious behaviour or fraud.

In 2022, a Dutch court issued a landmark ruling in favour of Bunq, the EU’s second-largest neobank, after it sued the country’s central bank for banning it from using AI and machine learning to conduct money-laundering checks.

“To a certain extent, there’s little room for caution,” said Ali Niknam, Bunq’s founder and CEO. “Either you onboard AI or get left behind.”

He said using AI to monitor payments helped to flag fraudulent transactions 2.5 times better than “traditional banking methods”. Elsewhere, Bunq claims its AI assistant Finn has answered more than a million user questions with a 70 per cent resolution rate.

“At Bunq, we expect to automate up to 90 per cent of our operations with the help of AI,” Niknam added. “Most importantly, we’re using AI to make life easy for our users.”

At the Trades Union Congress’ annual conference later this month, employee unions Unite and Accord, which include banking workers, plan to urge banks, insurers and accounting firms to fund a “major” programme to reskill millions of staff whose jobs could be impacted by AI.

They have cited a June report from Citigroup, which said around 54 per cent of banking jobs had a high potential to be automated – more than any other sector – while a further 12 per cent of roles could be augmented with AI.

Citi, itself one of Wall Street’s many adopters of the technology, added that AI could add $170m (£128.5bn) to banks’ profits by 2028.

But the shifting sands won’t necessarily cause a sector-wide drop in headcount. Firms are expected to hire in new areas like AI management and compliance to ensure the technology is being used properly and in accordance with regulations.

“Finn is creating jobs at Bunq, not the other way around,” Niknam said. “While generative AI is the backbone of our rapid growth and scaling across Europe, we’re able to actively hire across a variety of fields.”

Hirt said AI would “remain a key driver of progress” in financial technology in the coming years, with “transformative potential” in areas from predictive analytics to enhancing regulatory compliance.