Wealth and asset managers entering ‘second wave’ of AI adoption

Wealth and asset management firms are entering a second wave of artificial intelligence (AI) technology adoption, but poor data, legacy systems, and resistance from leadership remain obstacles to transformation, according to Softwire.

The consultancy noted that wealth and asset managers were shifting their AI focus from back-office automation to higher-value uses, such as risk modelling, investment research, and software development.

However, many were still struggling to build the data, technology, and organisational foundations required to deploy these higher-value use cases at scale.

First and second wave

Softwire’s research found that firms were already using AI to improve existing workflows, with back-office process automation the most common use (49 per cent), followed by client-facing AI tools (46 per cent), automated data extraction (43 per cent), and fraud detection (42 per cent).

Many were moving beyond this initial phase to a second wave of AI implementation, and adoption of AI-assisted software development, investment research, and risk modelling were each identified as the next AI focus by 33 per cent of respondents.

Just 21 per cent cited back-office automation as their next docus, which Softwire said suggested many firms were seeing it as work that was already underway.

“Asset and wealth managers are moving beyond simply using AI to automate processes and extract data faster, building the trust to tackle more complex use cases,” said Softwire director of financial services and insurance, Sean Judge.

“The next phase brings it closer to risk, research, and the decisions that shape performance.

“That switch is significant. Newer models are making more complex knowledge work possible, but higher-value use cases place a much more demanding test on the data, technology, governance and specialist capacity supporting them.”

Barriers to transformation

Softwire’s survey showed that while 98 per cent of firms had implemented at least one AI proof of concept or pilot into production over the past 12 months, just 12 per cent said AI had truly transformed their business.

Furthermore, 81 per cent of respondents said poor data and legacy systems were limiting their AI progress, while 44 per cent felt their data did not have the quality, lineage and traceability needed for regulatory confidence.

Over two fifths (42 per cent) said improved legacy modernisation would help them realise value from AI.

However, there were several barriers to achieving this, with 41 per cent citing the risk of disrupting critical operations as an obstacle and 40 per cent pointed to the challenge of integrating new technology with existing systems and data.

Other barriers included skills and capacity (38 per cent), cost and budget constraints (35 per cent), and regulatory or compliance concerns (34 per cent).

Over a third (37 per cent) identified leadership buy-in or cultural resistance as barriers to scaling AI, and 86 per cent expected to work with external partners in some form to ensure additional capacity and support that can connect technology delivery with business outcomes.

Softwire also highlighted a potential slowdown in client-facing AI, as while client-facing tools were the second most implemented AI use case, only 28 per cent cited them as their next focus.

“Client-facing AI is where the confidence test becomes much harder,” said Judge.

“Back-office automation can often be kept within a narrower process, but client-facing AI depends on more of the organisation being ready at the same time.

“In a regulated sector, it is not enough for the technology to be impressive. It has to be reliable, explainable and built on data the organisation can trust. That means strong data foundations, modern architecture and organisational buy-in all have to come together.

“That is why client-facing AI may take longer to scale, even as the models themselves become more capable. Firms will also need to build consumer trust by being transparent about when and how AI is being used.”

Judge added that while the research showed a sector with no shortage of AI ambition, there was a clear gap between experimentation and transformation.

“Firms are targeting more valuable use cases, yet fragmented data, legacy systems, operational risk and limited delivery capacity are making them harder to scale,” he said.

“The priority now should be to focus on a small number of high-value use cases, modernise the data and systems they depend on, and build clear business ownership around delivery. That is how firms can reduce risk, prove value and move AI beyond the pilot stage.”

---