ACA Group’s latest research report suggests buy-side AI adoption progressing cautiously from personal productivity desktop tools towards defined compliance and operations use cases.
Based on 201 qualified respondents across US-based investment management firms, the report examines AI use across 20 compliance and operations workflows in a market segment where AI exposure is widespread, but operational depth remains limited: the central finding is that the average firm is using AI in fewer than two of those 20 workflows, while 36% report no AI usage across these function.
Compliance Leads Adoption
Compliance is the strongest area of buy-side AI adoption. Workflows are often text-heavy, review-led and document-intensive, including surveillance, marketing material review, compliance programme administration, policy work, testing and monitoring. These areas suit current AI capabilities because they involve classification, summarisation, drafting, comparison, anomaly detection and review support. They also allow firms to introduce AI without changing core transaction processing or books-and-records infrastructure.
The strongest early use cases are therefore those where AI can improve the efficiency and consistency of human review. In eComms surveillance, AI can support triage and alert enrichment. In marketing review, it can help identify missing disclosures, inconsistent claims or policy breaches. In compliance administration, it can assist with documentation, workflow support and knowledge retrieval.
Desktop AI for personal productivity is the widest reported use case, with 84% of respondents using tools such as Microsoft Copilot, ChatGPT, Claude and Gemini. As an entry point, these tools can lower cultural resistance and help users understand where AI may add value.
Compliance testing stands out as a likely next phase of adoption. ACA found that 52% of respondents identified compliance testing and monitoring as an area where they want to see AI used next. That is significant because testing is a labour-intensive control function and a key source of assurance for boards, senior managers and regulators.
AI can support testing by expanding sample coverage, identifying patterns across larger datasets, preparing evidence packs, comparing activity against policy requirements and directing human reviewers to higher-risk items. This has practical value in firms where compliance teams face rising obligations, tight resources and increasing pressure to evidence the effectiveness of controls.
The challenge is defensibility. Testing is not a low-risk productivity task. Firms need to explain why an item was selected, how the AI output was assessed, what evidence was used, who reviewed the result and how issues were escalated.
Operations Takes Longer
Areas such as reconciliation, pricing quality control, market data quality, trade confirmation and net asset value validation are attractive candidates for AI because they involve data-heavy processes, exception handling and repeatable controls.
These workflows present tougher demands. They depend on order management systems, portfolio management systems, accounting platforms, data feeds and downstream reporting processes. Applying AI in these areas requires strong data foundations, lineage, explainability, exception management and evidence capture. The use cases are compelling, but the risk of disruption is higher.
Firms need to select use cases where AI can deliver measurable value, where the control environment can be defined and where human oversight can be preserved. Broad AI initiatives risk staying at the experimentation stage. Targeted deployment in high-impact workflows is more likely to build momentum.
Firms need tools that integrate AI capability with policy, oversight, case management, evidence-based testing, escalation and reporting. They also need governance features that support model oversight, permissioning and auditability.
The Buy-Side Position
Against the wider capital markets ecosystem, buy-side AI adoption appears more selective and cautious. Sell-side firms, market infrastructures, data providers and technology vendors are applying AI across larger-scale environments, including regulatory intelligence, surveillance, software engineering, client service, data management and operational resilience.
Many investment managers operate with lean teams and rely on third-party services and platforms, which helps explain the measured pace of adoption captured in ACA’s survey. As AI moves deeper into buy-side compliance and operations, progress is likely to come from targeted workflows where firms can show clear value, preserve oversight and build the evidence needed to scale.
Subscribe to our newsletter
