- A global survey finds 73% of asset and wealth managers say AI is critical, but most report only modest ROI and about 12% see zero or negative returns.
- Adoption is accelerating: 77% have an AI roadmap, ~71% expect to deploy generative AI within three years, and agentic AI remains rare today but is on near-term agendas.
- AI use spans back-, middle-, and front-office work (e.g., process automation, compliance monitoring, and client support), yet data quality, fragmented systems, culture, and regulatory uncertainty constrain scale.
- AI leaders capture outsized value by tightly linking use cases to business goals and investing in cloud infrastructure, governance and risk controls, and AI-skilled talent.
Read More
The Grant Thornton / ThoughtLab global survey of 500 senior executives in Q3 2025 provides a timely, wide-angle view of where the asset and wealth management (A&WM) industry stands with AI—but also exposes a widening gap between ambition and execution. While nearly three-quarters of leaders believe AI is critical, fewer report strong ROI: only modest gains for most, and roughly 12% have seen negative or zero returns.
Organizations are making real investments: 77% now have formal AI strategy and roadmaps; around 71% plan adoption of generative AI within three years; less than 10% currently use agentic AI with 18% planning to do so. Usage is spreading through business lines: back-office (code, processes, custody), middle office (compliance, regulatory & tax monitoring, data security), and front office (customer analysis, conversational support).
However, critical impediments persist: cultural resistance (slow-moving cultures), lack of quality data (incomplete internal processes for cleaning, normalizing, tagging), fragmented tech stacks, and regulatory ambiguity. These factors limit adoption scale, effectiveness, and trust in AI-generated or AI-supported decisions.
Comparative studies reinforce these findings. An EY survey of 100 A&WM firms found 95% have scaled generative AI into multiple use cases, 78% are exploring agentic AI—mirroring Grant Thornton’s findings—but only about a quarter are seeing substantial impact. Other reports (e.g., from KPMG) observe movement along the maturity curve, with more firms in the “developmental” phase and back-office leading the way.
Strategic implications for firms include:
- Those without strong alignment of AI strategy to business goals risk investing in low-value or low-impact use cases — experiments that don’t scale.
- Data foundations will increasingly become a competitive differentiator. Firms that invest early in cleaning, tagging, normalizing internal and external data will enable more reliable AI, faster deployment, and stronger regulatory resilience.
- Regulatory and governance frameworks will shape winners. As agentic AI and GenAI increase in use, firms able to responsibly build transparent, auditable mechanisms will attract trust from clients, regulators, and partners.
- Talent and culture are central. Upskilling, role-based training, leadership buy-in, and moving from pilots to production matter more than purely technological capabilities in determining AI ROI.
Open questions to monitor:
- How will regulatory regimes evolve to address liability, model bias, explainability, especially for agentic AI?
- What kinds of ROI (cost savings vs revenue uplift) will dominate in different asset manager segments (e.g. size, region, strategy)?
- How rapidly can firms move from generative to fully agentic AI, and what business structures will facilitate that safely?
- Will data quality initiatives lag technology tools, and how will firms bridge gaps? Which external data sources or partnerships will matter?
Supporting Notes
- Nearly three-fourths (73%) of asset management industry executives say AI is critical to their organization’s future.
- 77% of firms have an effective AI strategy and roadmap in place.
- 71% are planning to adopt generative AI (GenAI) within three years; fewer than 10% currently using agentic AI; 18% plan to use agentic AI over next three years.
- Usage by business function: back office (code development 46%, business processes 42%, custody services 39%), middle office (57% regulatory/tax monitoring, 52% data security), front office (59% customer analysis, 58% conversational support, 54% self-service portals).
- Cultural and technological hurdles are major barriers: over half of respondents cite slow-moving culture and lack of access to quality data.
- Roughly two-thirds report only modest ROI; about 12% have no returns or negative results.
- EY’s survey: 95% of firms scaled GenAI to multiple use cases; 78% exploring agentic AI; but only a bit more than 25% report substantial business impact.
- KPMG: Firms moving from conceptual to developmental phase of AI maturity; back-office functions currently leading adoption.