Shield, a global communications risk management company purpose-built for financial institutions, has published new research showing that banks are ramping up spending on AI-driven surveillance yet remain unable to translate that investment into operational change.
The study, commissioned by Shield and produced by 1LoD, found that more than 90% of institutions are still operating in alert-heavy environments, with false positives ranked as the sector’s most significant surveillance challenge, a position unchanged since 1LoD’s previous benchmarking survey in 2024.
False positives were rated highly significant by 52% of the firms surveyed. Budget and resourcing constraints followed at 41%, while limited access to quality data and legacy or outdated systems were each cited by 37% of respondents. Regulatory pressure, by contrast, was flagged as highly significant by just 7% of firms, a gap the report says shows that execution, rather than regulatory uncertainty, is now the main obstacle facing most institutions.
Shield’s own deployment figures, referenced in the report and based on evaluations covering millions of communications, point to roughly a threefold cut in alert volume relative to legacy systems.
The company says this has come alongside accuracy gains of up to 44% and around three times as many actionable escalations, suggesting that a smaller volume of better-targeted alerts can sharpen risk detection rather than weaken it.
1LoD is a specialised conference and intelligence provider focused on non-financial risk and compliance across the banking sector, and produced the underlying report using responses gathered from senior surveillance and compliance professionals at financial institutions.
The report also sets out what distinguishes compliance organisations that have successfully operationalised AI from those still struggling to do so. Leading firms are said to be shifting away from static, rules-based detection towards adaptive models built on behavioural patterns, consolidating fragmented data and workflows into unified platforms, and linking surveillance spending directly to measurable cuts in manual review work.
Looking ahead, the report concludes that surveillance is heading towards a model built around integrated data, context-driven detection and reduced operational noise, with outcomes that can be measured.
Institutions that close the gap between the technology available to them and how it is actually used, it states, stand to cut costs while strengthening genuine risk detection.
Shield co-founder and CEO Shiran Weitzman said, “AI ambition is everywhere right now. What’s missing is execution. Investment is accelerating, but the fundamentals still matter. AI cannot deliver its full potential when surveillance is constrained by fragmented data, legacy infrastructure, and alert-heavy operating models. Closing that gap is what will separate firms that simply deploy AI from those that fundamentally improve how risk is detected and understood.”
Copyright © 2026 FinTech Global
