The financial services industry is accelerating towards an AI-first future, but so far, enthusiasm and execution sit far apart. When we surveyed 1,000 leaders across financial services and fintech, 92% told us they are ready for agentic AI, but only 36% have a funded strategy to guide the shift, highlighting a clear readiness divide. The aspiration is there, but the foundations are not yet in place.
Now, they face the challenge of closing this gap without falling behind early adopters. Here, our experts explore what might be causing it, and how to move from readiness to execution.
Why the confidence-capability gap matters
Already, 16% of financial services and fintech leaders would describe their organisations as AI-native, while a further 36% expect to achieve this by the end of 2026. But without a funded, enterprise-wide plan in place, this will prove challenging.
So, what are the top barriers to achieving this AI adoption?
-
Cybersecurity and data privacy risks (31%)
-
Data quality, availability or integration challenges (29%)
-
High costs of implementation and scaling (28%)
-
Regulatory uncertainty and compliance complexity (25%)
-
Concerns around trust, transparency and explainability (25%)
-
Legacy systems (24%)
-
Cultural resistance and talent shortages (21%)
With no single overwhelming obstacle, organisations must think carefully about what is causing their own barrier to adoption. Without this, it will be challenging to bridge it.
Identifying true readiness
While challenges such as cybersecurity and legacy systems can slow pace, for many organisations, there can also be a gap between perceived readiness and true readiness. When organisations say they are prepared for AI-native transformation, they may simply mean they understand the need and have begun exploring what it might look like. That awareness is necessary but not sufficient. Instead, true readiness should be measured across multiple dimensions:
-
Organisations should begin with strategic importance: is this central to your business or side-of-desk activity? Do you have financial commitment and executive sponsorship, or are people fitting AI work around other priorities?
-
Leaders should also consider their organisation’s foundations. Data quality, skills, culture and technology platforms all need to be in place before autonomous systems can operate effectively.
-
Assessing impact is crucial. What value are you actually achieving with AI today? Use cases that are in production matter much more than pilots, and measurable outcomes matter more than promising experiments. A gap between activity and impact showswhere perceived readiness diverges from demonstrated capability.
-
Finally, institutional knowledge should be measured. Understanding how autonomous systems integrate with your specific business context, how to govern them effectively and how to measure success in ways that go beyond traditional metrics cannot be rushed or purchased. This learning comes from hands-on work. Organisations that delay starting this work in pursuit of perfect clarity or complete buy-in are falling behind competitors who are learning by doing. In practice, much of this institutional knowledge is implicit, trapped in individuals rather than documented in systems or processes, which can cause automation efforts to fail as edge cases, workarounds and informal decision-making are not designed for. Before introducing autonomous systems, examine core processes critically with this in mind, identifying where they break down in reality, not just how they are meant to operate on paper.
It's important to note that the readiness perception gap often differs across departments. Technical teams may believe platforms and infrastructure are excellent, while business stakeholders view them as inadequate. Data teams may be confident about quality and availability, while those trying to use that data for decision-making find it incomplete or inaccessible. This can cause delayed decision-making, friction or a lack of stakeholder buy-in.
Agreement matters even when the current state is poor. If different parts of the organisation hold fundamentally different views about readiness, those views need to be surfaced and resolved before transformation can progress.
Identifying AI investment priorities
Part of the challenge in building a strong, fundable strategy is that AI investment is not a decision made in isolation. It spans multiple categories, each requiring different approaches to evaluation and different expectations for returns.
For example, an organisation may consider the following categories:
- Productivity tools
These tools, like ChatGPT Enterprise, roll out broadly to enable gains across a range of everyday tasks. Value is diffused and often measured retrospectively. These tools also serve as the primary mechanism to deliver the rising tide of AI capabilities. Building the business case requires accepting that precise ROI may be difficult to quantify upfront. - Systems uplift
This comes through AI features being added to existing platforms. Salesforce, ServiceNow and other enterprise tools are embedding increasingly sophisticated AI capabilities. Organisations that maintain current versions get access to new capabilities without significant additional investment. The value compounds over time but is tied to the broader platform relationship.
- Business-as-usual use cases
In these cases, AI can deliver tactical value without fundamentally changing the business model. Automating document ingestion, accelerating analysis or reducing manual errors in existing processes creates measurable efficiency gains. These are often the easiest investments to fund because they fit within traditional ROI frameworks. They also build the skills and confidence needed for more transformative work.
- Reinvention
These investments redesign processes or business models around what AI makes possible. These are the hardest to fund because traditional ROI metrics often fail to capture the value. The benefit may be preventing competitive destruction rather than generating direct returns, enabling a new business model that does not exist yet or building institutional knowledge that becomes a strategic asset. - Automation
This category focuses on using AI and agentic systems to automate existing processes, often acting as an evolution of traditional robotic process automation (RPA). Rather than changing the business model, organisations use AI to speed up workflows, reduce manual effort and manage process variability.
Examples include automating customer onboarding, handling payment exceptions or coordinating compliance and service workflows. These initiatives typically deliver clear efficiency gains and are easier to fund within traditional ROI models, while also building experience with AI-driven orchestration.
Leaders need to be intentional about how they balance investment across these categories. Focusing only on areas with clear ROI leaves you optimising existing models whilst competitors build new ones. Investing only in transformation without securing quick wins makes it difficult to maintain momentum and stakeholder confidence.
Moving from aspirational strategy to a funded roadmap
Closing the confidence-capability gap requires moving from high-level conviction to specific, clearly fundable initiatives that build capability systematically. To achieve this, leaders need to identify a clear vision, prioritise their investments carefully and look beyond traditional ROI to capability building and strategic resilience.
We often encourage customers to think about investment as becoming an AI company rather than implementing AI products. The distinction matters; products become commoditised, but institutional knowledge, cultural adaptation and organisational capability compound as strategic assets.
The gap between the prepared and funded is not down to a failure of vision, but a practical challenge of moving from strategic conviction to operational capability. The organisations closing that gap are not waiting for perfect clarity. They are building institutional knowledge, securing incremental funding and learning by doing.
Discover more insights from our research in our latest report.