New research, based on interviews with 400 senior executives across major financial institutions, has found that as the sector tries to scale its AI implementation, siloed data is the biggest barrier to wider adoption. In spite of this, three-quarters of businesses are expecting to scale from their piloting phase, as the technology continues to evolve.
Generative artificial intelligence (GenAI) is a subset of AI technology, using generative models to produce text, images, videos, or other forms of data. While proponents of the technology argue this offered unparalleled speed and efficiency in processing complex data, such models are famously prone to ‘hallucinations’, and often generate inaccurate responses to queries. This means that while the financial services sector has been no stranger to spending on the technology, research has previously shown that levels of maturity remain low.
Now, a study from innovation consulting firm Valtech has explored how the financial services space remains slow in its adoption – and what the key barriers are. And the researchers found that attitudes on reliability are actually shifting.
Source: Valtech
At present, 44% of FS executives cite improved reliability/risk management as their main motivation for AI investment. Meanwhile, a further 17% of respondents cite an ‘improved speed-to-market’ as the primary selling point, while 17% see AI as an opportunity to drive revenue growth.
Meanwhile, almost half of firms in the sector plan to expand their data and AI teams. Wealth managers leading the way, with 61% suggesting this kind of investment is heading the way of their firm, ahead of 43% of banks, and 41% of insurers. This uptake is highest in Switzerland, with 62% of leaders there planning to grow data and AI teams – but even in the more cautious UK, that portion is approaching half, at 48%.
Even amid this heightened enthusiasm, however, Valtech found that 77% of AI initiatives were only just scaling from ‘pilot’ to ‘production’ – despite investments in the sector dating back for more than three years now. So what is causing the delay? According to the researchers, the technology may have outpaced the structures designed to support it – leaving gaps in implementation.
Source: Valtech
In particular, old operational structures are seeing AI implementations hamstrung. The key challenge this causes is unclear data ownership, with 48% of leaders saying siloed data is the key barrier to scaling AI within their organisation. Meanwhile, 16% of respondents blame governance friction and slow approvals, 12% say legacy systems are the biggest hurdle and one in ten cite fear of ‘backing the wrong technology’ as the key rationale for delays in transformation.
In the year ahead, many companies believe their investment in data and AI talent may help to contend with this – with embedded technical expertise helping to smooth out implementations. This may overlook the need for more extensive operations shifts, though, as just 7% are looking to move to cross-functional, outcome-led models – and without dealing with siloes, that may still put the brakes on progress for some projects.
Gary Ellison, head of data and AI for Europe at Valtech, commented, “Soft barriers remain a friction point, and a cultural consideration for financial services leaders. Our research shows that different regions are having notably different success rates when it comes to rollout of AI projects. Organisations must focus on identifying these barriers, both technological and cultural, in a bid to see projects through from pilot to production.”
Source link
#Financial #services #firms #find #siloed #data #standing #rollout
