Kathy Gibson reports – The financial services industry is under pressure to quickly adopt artificial intelligence (AI), but needs to ensure that it does so responsibly.
There are huge opportunities for the industry to increase innovation and profitability, but there is still a big gap between interest and reality.
“This is the time to reimagine the way things ae done,” says Indranil Bandyopadhyay, principal and researcher at Forrester.
“AI is not a sledgehammer: you need to use it responsibility, and in the right places,” he adds. “Take inspiration, learn from it, and use it in your company’s unique context.”
According to Forrester’s report, The State Of GenAI In Financial Services, 2024, 58% of global business and technology professionals at FS companies have GenAI-enabled use cases in production, while 14% have implemented pilots. Additionally, 44% cite productivity and revenue growth as the main benefits of utilising the technology. Forrester further finds that technology departments in FS companies command “the lion’s share” of genAI exploration (69%), while the business is responsible for only 13%.
Current trends in financial services aim to offer customers services more personalised and seamless services.
Financial services should be invisible. “You need to be part of the client’s needs and journey they are in. whatever financial services offers needs to be a sidebar issue.”
They need to be connected, collaborating with other players to offer invisible services to customers. “Our research shows that you cannot do it on your own,” Bandyopadhyay says.
It goes without saying the financial services needs to be insights-driven, keeping a finger on the pulse of what customers are doing. At the same time, being purposeful is key.
The typical customer today is evolving new channel behaviours, and trusting more in startups and platforms.
At the same time, new competitors are entering the market, from anywhere in the world and using distribution platforms. This limits players’ ability to differentiate on brand.
High operating costs continue to plague financial services companies. At the same time there is weak economic growth coupled with a cost-of-living crisis.
A lot of regulations accompany the evolution of open banking into open data and open finance. Other commitments like ESG and sustainability add to red tape. These all contributed to challenges for operational resilience, Bandyopadhyay says.
Indranil Bandyopadhyay, principal and researcher at Forrester.
The potential annual value of AI services could be in the region of $1-trillion by 2030 – and financial services could benefit the most from automation and augmentation, according to the World Economic Forum.
The supply of AI into the financial services industry is expected to rise from $38,36-billion today to $190,33-billion by 2030.
Forrester research indicates that 52% of financial services companies are promoting AI to the highest level of investment within the next 12 months.
The time is right for investing in AI, Bandyopadhyay says. Not only are rapid technological advancements making it possible, but the industry also finds itself with an adundance of data. “What are we doing about this data?”
“You need to understand the customer – proactively, not just after they speak to you; you need to understand their requirements; the backend needs to be in place; and you need to have an idea of how you use the data.”
Forrester shares that in its September 2023 Artificial Intelligence Pulse Survey, 94% of participants indicated that their organisation’s internal data was of high quality and ready to use for genAI use cases, while 75% said they had rules in place to govern how their employees used public genAI tools for work.
However, analysts caution that “the hype surrounding genAI and the Dunning-Kruger effect are likely behind this confidence”, pointing out that respondents list governance and data quality as among the top impediments to rolling out consumer digital experiences in FS.
Financial services companies struggle to use the data they have (34%), can’t get access to it (33%), and can’t process it fast enough (33%), for digital experiences.
And, despite concerns about privacy and data protection, nearly half of the industry (47%) has not yet actively done such assessments.
That financial services companies are under competitive pressure goes without saying, along with heightened customer expectations. And companies are under economic pressure to do more with less.
AI can help companies to improve customer experience and open up new revenue streams.
Not only can it increase efficiency and reduce costs, it can help to enhance risk management through better fraud detection, credit risk management and regulatory compliance.
However, the financial services industry has some pretty unrealistic ideas about AI, Bandyopadhyay says: 25% think they can achieve an ROI of 51% to 75% over the next couple of years; 26% aim for an ROI of 26% to 50%; 19% think they will have an ROI of 11% to 25%; and 7% expect an ROI of 6% to 10%.
More unrealistically, 18% believe they will achieve an ROI of 76% to 100%, and 3% think there will be an ROI of greater than 100%.
“We need to be realistic,” says Bandyopadhyay. “My prediction is that less the 5% will make tangible ROI in 2025.”
Financial services companies must beware of the pitfalls including data quality and bias, black-box model, ethical and regulatory concerns, operational and cybersecurity risks, reputational risk, talent and expertise gap, and realistic ROIs.
Most companies are experimenting with AI, but very few of them are taking the leap into operational systems today, Bandyopadhyay says.
There are a number of barriers to implementing AI in financial services:
- Data privacy and security concerns – 32%
- Difficulty integrating with existing infrastructure – 23%
- Data infrastructure – 22%
- Lack of technical skills – 22%
- Lack of understanding of where to apply AI – 20%
- Trust in AI systems – 20%
- Lack of clear AI strategy – 19%
Perhaps the biggest issue is trust, with 30% of customers concerned over the use of AI, along with a 27% increase in AI-generation malicious content affecting brands and shareholders taking a bigger interest in responsible AI efforts (26%).
Indeed, a significant 25% of innovation has been halted because of trust issues.
Forrester’s Predictions 2025: Insurance report anticipates an 8% increase in tech spending across the insurance industry in 2025, with AI a key focus. However, the company says despite the excitement surrounding AI, fewer than 5% of insurers are expected to see direct, tangible gains from the technology this year due to such barriers as legacy systems, a lack of AI talent, and integration challenges.
In implementing AI, Bandyopadhyay reiterates the AI maturity moves from assisted, to augmented, to automated, to autonomous systems.
Companies should prioritise projects, examining each AI use case through the lens of business, governance, data science and technology stakeholders. There’s currently a skewed focus on GenAI exploration that’s mostly led by technology teams (69%) rather than business units, which could create risky gaps.
It’s important to follow some kind of activation framework, Bandyopadhyay adds. And, if use cases don’t work, companies shouldn’t be afraid to drop them.
The payoff journey may take longer than expected, he adds. Low-impact use cases may h=show quick ROI, but high-impact user cases will typically take longer.
Bandyopadhyay points out that AI will work in conjunction with other, complementary technologies like edge computing, neuromorphic computing, computer vision, AI-optimised hardware, and physical robots.
Looking to 2025, Forrester predicts that most enterprises fixated on AI RIO will scale back prematurely.
Speaking to the importance of data, 30% more large firm CIOs will being CDOs into the IT fold, and 40% of highly regulated enterprises will combine data and ai governance.
For 50% of use cases, the AI pendulum is expected to swing back to predictive AI, with three out of four firms that build aspirational agentic AI models set to fail.
Bandyopadhyay’s advice for financial services companies is to create an AI strategy that aligns to the business strategy.
“Prepare your data for AI,” he adds. “If you don’t have the fuel, you cannot drive any car.”
He suggests companies identify a high-impact, low-risk project to start with, and to create a regulatory sandbox. Finally, it’s important to measure and iterate.