As with any previous wave of technology-powered change, it can be difficult to evaluate the real-world impact of AI amidst the hype and speculation over what it may deliver in the future. It was a pleasure to attend the AWS Financial Services Analyst Day and the Symposium that followed. It was a great level-set in that respect, with the opportunity to discuss GenAI and agentic experiences with practitioners across the industry.

Taking together the insights shared in both the sessions and side conversations, I came away with three core messages about how AI is changing the banking industry.

1. Beyond efficiency: Enhancing customer engagement through AI

Most of the banks implementing GenAI and Agentic technologies today began their journey with the primary aim of delivering process efficiencies. Given the broad range of opportunities and the relative ease of getting business case approval for anything that will lower costs, this makes a huge amount of sense. Against this backdrop though, it was notable to see a shift in emphasis towards the role that AI investments are playing in driving stronger customer engagement.

US Bank’s contact centre transformation project neatly illustrates this point. What started as a technology-led initiative to consolidate operations across the bank’s 100+ contact centres onto Amazon Connect Customer turned into a broader reimagining of the wider customer journey.

The bank quickly realised that the real win was not reducing average handling or call waiting times (which it achieved), but in enabling more seamless transitions across channels and by enabling agents to resolve a higher proportion of complex interactions on the first call. Capabilities such as call summarisation and real-time knowledge assistance were the key building blocks here.

This wasn’t the only example, and we heard from other institutions using AI-driven meeting summarisation and knowledge retrieval to improve the customer experience (and productivity) of investment advisors.

The clear message coming through is the importance of factoring (and tracking) the impact on customer engagement from any AI investment that touches end-user experiences or processes.

2. Preparing for the next customer segment: AI Agents

The idea of AI agents acting autonomously to conduct transactions (including making payments) is not new to the industry, but it’s also clear that this has moved rapidly from theory to practice.

Stripe, in collaboration with AWS, has begun to build infrastructure to support Agentic Commerce. This includes enabling agent-enabled payments and wallet capabilities.

For a merchant, the implications for product design are significant:

• Discoverability becomes critical. Agents do not browse for products; they query. If products are not structured in a way that makes them easily discoverable and comparable by AI agents, they effectively disappear.
• Friction becomes existential. Human users may tolerate complexity or imperfect flows, especially if they have a brand attachment to the merchant. Agents will not. If checkout or authentication introduces delays or ambiguity, the transaction will simply move elsewhere.
• Identity and trust become central challenges. Banks and merchants must be able to distinguish between ‘good’ agents and potential bad actors, understand who an agent represents, and enforce constraints on behaviour.

Alongside digital commerce, AWS is also bringing new functionality to its Agentic toolkit. One of the big announcements at the event was the launch of Amazon Bedrock AgentCore Payments. This brings the ability for autonomous agents to react independently to HTTP 402 ‘Payment Required’ status codes to access pay-per-use digital resources (such as APIs, content, and other services) where needed to complete a workflow.

The takeaway for banks is not just about adopting new rails. It’s about preparing for a world where customers will increasingly engage AI Agents to transact on their behalf, whether that’s to make payments or to handle their finances.

3. Practical learnings: There is more to AI than technology

If the main messages from the event highlighted what we’re seeing in terms of product and service innovation, the underlying theme was about how to deliver successful execution. A consistent set of lessons is emerging here.

The importance of data architecture and governance is critical. Ultimately, the institutions that have invested in cloud migration, standardised data models, and governed data environments are the ones moving most quickly. In other words, AI will not fix data problems. It will amplify them.

Another key message is that the real transformation is led by changes to processes and workflows, not through applying technology alone. Automating a single step in a fragmented process can only yield marginal gains. Redesigning the process itself, by removing handoffs, eliminating duplication, and embedding AI within the flow, delivers step-change improvements.

The final piece is the importance of people. AI is changing how work is done, who does it, and the skills that are required. Scaling AI, therefore, requires sustained investment in training, enablement, and cultural change. Without this, even well-designed initiatives will struggle to gain traction.

It is also critically important to keep in mind the needs and preferences of the customer. In many use cases, implementing AI is as much a customer experience play as it is a workforce transformation or tech rollout, and needs to be framed as such.

Final thought

While concrete action will always lag the hype when it comes to any technology, it’s clear that many banks are rapidly adopting AI to reshape the way they serve and engage customers. 

Indeed, RFI Global’s research demonstrates there is certainly plenty of appetite out there for AI. Around 29% of customers are looking to their bank to provide AI-enabled financial solutions, while 8% are already using GenAI tools for financial advice and management.

These themes, and others, are high on our agenda for 2026, so contact us if you would like to learn more.