Financial firms struggle to scale AI despite strong returns

Riverbed has published survey findings suggesting most financial services firms still struggle to move artificial intelligence projects from pilots to enterprise-wide deployment, despite strong stated confidence in their strategies and the returns from AIOps investments.

The report, based on a global survey of IT and business decision-makers, found that 88% of AI projects in financial services were not deployed enterprise-wide. In addition, 62% of AI initiatives in the sector remained in pilot or development.

The data points to a gap between intent and execution. Nearly two-thirds of respondents reported high confidence in their AI strategy, and 89% reported that AIOps returns met or exceeded expectations.

Even so, only 12% reported that their AI initiatives had achieved full enterprise-wide deployment, and only 40% felt fully prepared to operationalise their AI strategy.

Data readiness

Data quality emerged as a central issue. The survey found that 92% of decision-makers agreed that improving data quality was critical to AI success, yet only 43% were fully confident in the accuracy and completeness of their organisation's data.

The findings contribute to a broader debate in financial services on governance and oversight of AI systems, particularly as firms increase the number of production deployments. In the UK, the Financial Conduct Authority has launched a review of how AI could reshape retail financial markets through 2030.

Some banks are beginning to quantify value from scaled deployments. Lloyds Banking Group has reported that generative AI delivered approximately GBP £50 million in value in 2025 and expects this figure to exceed GBP £100 million in 2026 as it expands its use across operations and customer experience.

Insurance supervisors have also signalled a growing focus on technology change. The International Association of Insurance Supervisors plans to provide a platform for knowledge sharing on emerging digital innovation, including AI.

Tool sprawl

Operational complexity featured heavily in the findings. Financial services organisations reported using an average of 13 observability tools from nine vendors to monitor AI deployments. Fragmented tooling can make it harder to correlate performance and reliability issues across applications, networks and user experience.

Almost all organisations surveyed were consolidating tools and vendors across IT operations (96%). Most respondents agreed that a unified observability platform would make it easier to identify and resolve operational issues (95%), and 95% reported considering new vendors as part of consolidation efforts.

Riverbed linked the consolidation trend to the needs of regulated environments, where firms tend to prioritise auditability and operational resilience. They also face pressure to maintain the stability of digital services as AI workloads grow and transaction volumes remain high.

UC performance

The report also highlighted the impact of unified communications performance on IT operations. Respondents said employees spent 41% of their working week using unified communications tools. Only 47% reported being very satisfied with their performance, while 44% reported experiencing regular issues affecting video calls, messaging platforms, and collaborative workspaces.

UC-related incidents accounted for 16% of IT tickets in the sector. The average reported resolution time was 41 minutes, and nearly one in five took more than an hour.

Open standards

The research indicated broad adoption of OpenTelemetry in financial services, with 92% of organisations already using the framework. Cross-domain correlation also featured strongly, with 96% calling it critical to their observability strategy.

Almost all respondents (99%) agreed that OpenTelemetry reduced vendor lock-in and increased flexibility. A further 97% viewed it as a foundation for future initiatives, such as AI-driven automation.

Networks and data movement

The findings suggest financial services firms are paying closer attention to where data is stored and how it moves across networks as AI use expands. The survey found 94% viewed AI data movement as important to their overall AI strategy; 37% described it as critical and foundational to how they design and execute AI.

Network performance and security also ranked highly, with 81% citing both as essential, Riverbed's highest level among the industries surveyed.

Looking ahead, 76% of financial services organisations reported plans to establish an AI data repository strategy by 2028. That points to further investment in governed architectures across public cloud, edge and co-location environments, alongside continued scrutiny from regulators and supervisors.

Riverbed Chief Marketing Officer Jim Gargan said the challenge is moving from experimentation to consistent production outcomes.

"Financial services organisations are among the most sophisticated and disciplined adopters of AI, and our research shows they're already seeing strong returns. However, the sector operates under unique pressures, including rigorous regulatory scrutiny, zero tolerance for downtime and a critical need for data accuracy. Success now depends on simplifying IT, consolidating observability tools and vendors, improving data quality, embracing open standards like OpenTelemetry, and ensuring network and application performance can support AI at scale. At Riverbed, we are actively supporting some of the world's largest financial services organisations as they bridge this gap and turn AI ambition into operational reality."

The survey polled 1,200 business decision-makers, IT leaders and technical specialists across seven countries and multiple industries, including financial services. Research was conducted by Coleman Parkes Research.