Last week I had the pleasure of attending two financial services industry analyst events in New York:
- HCLTech Financial Services Advisor and Analyst Day
- AWS Financial Services Analyst Summit, followed up by the Financial Services Symposium
Although very different players in the financial services ecosystem, there were similarities in how these vendors hosted these events.
- Focus on vertical industry analyst firms, plus some breadth firms with a small industry group
- Few slides, more fireside chats, and panel discussions with clients and partners
- Access to meet clients and partners, and discuss challenges and opportunities firsthand. It was a pleasure to meet with several bank executives who shared their journeys to digital transformation and AI.
Right on the back of that was the Red Hat Summit in Boston that I joined remotely. A broader enterprise view rather than an industry focus, but also with great customer testimonials. As always, I appreciate the invitation from key vendors to keep us at Celent informed and engaged in relevant product, partner, and client news – and the fact that a couple of these were back-to-back in lower Manhattan helped too!
Key impressions
AWS - When using Amazon Q Developer, one financial services firm experienced developer productivity improvements of 40% in a pilot. This was extrapolated to $120Mn in technology cost savings. Now that firm has trained over 1,500 developers as the program is scaled across the enterprise.
GenAI unlocked massive scale in discovery and creativity. What it didn’t have was memory, the ability to learn and apply contextual reasoning, or to dynamically create plans and execute tasks. Unlike GenAI and RPA, Agentic AI architectures bring all these elements together - where agents have specialized roles, and access to information and tools, to adaptively plan and execute tasks.
HCLTech - The head of platform transformation at a top five US bank spoke about their partnership with HCLTech and how their retail banking production operations is being transformed. He spoke about a vision for agentic AI systems to self-heal production environments. It’s a really interesting point. “Newsworthy” production issues are Severity 1 or 0 – where a production outage requires recovery of a business-critical system. Thankfully, these are relatively rare, but banks also have a long catalog of much lower priority incidents that should be fixed in future releases. Inevitably, many of these never get addressed and can come back to bite an organization in a bigger way. If agentic AI systems can heal and patch these “ankle biter” issues, they can reduce the likelihood of them escalating into major issues. It would also free developers and support teams to focus on high priority items.
Red Hat - Simply – AI anywhere (my words). That’s a powerful concept, especially in an era of increasingly complex digital sovereignty in certain parts of the world, and the need to address resilience and application portability requirements, such as the DORA* regulation in Europe. Understanding digital sovereignty and managing data and AI assets in today’s interconnected digital banking environment is very complex and requires consideration of both legal and architectural/infrastructure engineering impacts. The ability to build AI locally (effectively containerize it) and deploy anywhere – on-premise, hybrid or public cloud - gives banks flexibility in deploying AI across hybrid datacenter footprints in multiple jurisdictions, or with multiple cloud partners for resilience and portability.
In many ways, the excitement about Agentic AI in 2025 resembles that for GenAI in 2023. Remember that the first reaction of financial institutions was to ban ChatGPT? The next year saw a wild frenzy of imaginative use cases and pilots that had a pretty short runway to become production solutions. Now with the promise of agentic AI systems, buckle up for 2026!
