Mercer: 55% of asset managers now use AI in investment processes

TL;DR:

• 55% of 131 surveyed asset managers worldwide have integrated AI into at least one investment process, with a further 27% running pilots — leaving just 18% with no current AI in investing.
• 91% plan to increase AI usage in the next 12 months, but only 8% report measurable improvements in investment returns from current AI implementations.
• Data quality (cited by 69%) and regulatory/compliance concerns (59%) remain the two principal barriers, with Mercer characterising AI as still “a partner rather than a decision-maker”.

The Mercer report, How Artificial Intelligence is shaping asset management, gives the UK financial services industry one of the first credible cross-border benchmarks for AI penetration in active investing. The headline finding — adoption past the majority threshold for the first time — matters less than the productivity-versus-returns gap underneath it.

Context and Background

The pattern Mercer documents is now familiar across sectors: enthusiastic deployment, easy operational wins, harder investment outcomes. 73% of asset managers use AI for operational efficiency in existing teams, and 68% deploy it as an analytical “partner” alongside human portfolio managers. Only 5% have granted AI any autonomous or semi-autonomous trade or recommendation authority — a number that should reassure UK pension trustees and IFA gatekeepers concerned about model accountability.

The 8% figure for measurable return improvement is the more important data point. It aligns with the broader enterprise AI experience over the last 18 months: cost-out and productivity gains arrive quickly, while top-line or alpha-generating impact remains difficult to attribute. For UK asset management firms now writing their 2027 strategy plans, that suggests budget arguments built on “AI-driven outperformance” will struggle to find supporting evidence, whereas arguments built on operating-cost reduction sit on firmer ground.

Mercer global manager research leader Beverley Sharp pointed to data and regulatory friction as the binding constraints — both areas where UK-regulated managers face additional FCA scrutiny under the Consumer Duty and operational resilience regimes. Mercer also disclosed it has built its own AI-powered manager research tool to streamline due-diligence drafting, putting it among consultants commercialising AI tooling rather than just advising on it.

Looking Forward

The 91%-plan-to-expand figure means UK fund selectors should expect AI capability claims to become a standard line in 2026 fund manager pitches, particularly around research workflows and due-diligence acceleration. Resultsense expects the gap between adoption claims and demonstrable return impact to be the next major scrutiny point — for trustees, regulators and investment consultants alike — through the second half of 2026.