FinovateEurope 2025 focused on proving that AI agents could work. Returning to London on 10–11 February this year, the focus shifted to operationalizing agents safely, compliantly, and at scale inside regulated banks. The strongest demos emphasized governance, explainability, and integration into existing bank estates — reflecting the growing influence of regulators, risk, and compliance teams on fintech innovation. During the Finovate all-star session, I shared insights from my research on the future of digital banking experiences. However, across two days of demos and industry-stage conversations, a clear transition emerged: Experimentation is giving way to execution. The themes below capture what stood out most.
Agentic AI Is Moving From Experimentation To Governed Execution
Demoers — including Darwinium, Intuitech, Maisa, Neuralk-AI, and Sea.dev — showed agents performing real operational work, from document analysis and data validation to credit and fraud checks, workflow orchestration, and structured outputs for human review. What differentiated these demos wasn’t autonomy alone, but “controlled autonomy.” Vendors combined deterministic process logic with LLM‑based reasoning, which ensured full traceability to source data, and built in clear mechanisms for human inspection, validation, and override. Across multiple demos, teams repositioned humans from primary operators to validators and overseers of AI reasoning — signaling a fundamental shift in back‑office operating models. The emphasis on auditability, replayability, and hallucination resistance indicates that firms now engineer agentic AI to withstand regulatory scrutiny — not just deliver productivity gains. This closely aligns with Forrester’s AEGIS Framework, which frames agentic systems as inherently risky by design and calls for deterministic guardrails, observability, and human control in regulated environments.
Legacy Modernization Is Shifting From “Rip And Replace” To Explain, Map, And Evolve
A cluster of demos — including R34DY, Sea.dev, and Tweezr (Sea.dev and Tweezr both won Best of Show for the second year running) — focused on helping banks modernize complex legacy estates without wholesale system replacement. These solutions use a combination of deterministic analysis, knowledge graphs, and AI to document existing systems, map technical and business dependencies, unpack embedded business rules, and make legacy logic explainable to engineers, product teams, and regulators. Rather than abstracting legacy away, these demoers emphasized deep understanding to enable precise, low‑risk change. This reflects a growing recognition that regulatory pressure, cost constraints, and operational risk make incremental modernization — not greenfield rebuilds — the dominant path forward. According to Forrester’s 2025 data, developers in financial services most frequently cited upgrading and migrating code from older language versions and migrating code to newer languages as key use cases for automating with TuringBots.
Fraud And Identity Are Shifting From Point‑In‑Time Checks To Continuous Behavioral Intelligence
As our research shows, the financial services industry faces a growing need for more seamless, effective, and secure identity verification as fraud tactics become more sophisticated and regulatory requirements more complex. Finovate demoers showed that identity verification and fraud prevention are evolving beyond static onboarding and episodic step‑up authentication toward continuous, real-time assurance across the customer lifecycle. Vendors — such as Candour Identity, Darwinium, Elephant, and Keyless — demonstrated how behavioral biometrics, motion and interaction analysis, device intelligence, and deep identity graphs can be used to detect deepfakes, synthetic identities, account takeover, and stolen‑device fraud. These solutions collect signals passively and continuously, reducing reliance on intrusive challenges while improving detection once attackers pass initial verification.
Trust, Inclusion, And Explainability Are Being Engineered Into Core Systems
Across multiple demos, trust was treated not as a policy layer but as an engineering problem — designed into workflows, measured through data and behavior, and evidenced at scale. Serene, which earned Best of Show, demonstrated behavioral intelligence that continuously surfaces vulnerability signals and contextual insights to frontline staff, enabling earlier intervention and more appropriate customer support. Hagbad showed how community‑based platforms can build portable financial identities using behavioral data, allowing individuals without formal credit histories to establish trust and access services. Mifundo focused on cross‑border credit data portability, enabling fairer lending decisions for mobile consumers. FINTRAC highlighted tools that automatically document models, calculations, and decision logic, ensuring outcomes are explainable, auditable, and repeatable for regulators and internal oversight.
Embedded Finance Is Maturing Into Modular, Bank‑Controlled Infrastructure
Demoers like AAZZUR and Opentech showed how firms can increasingly deliver embedded finance as reusable, composable building blocks — cards, lending, instalments, payments, and value‑added services — embedded directly into merchant and platform journeys. Unlike earlier waves, control now clearly sits with regulated banks rather than third‑party intermediaries, allowing banks to distribute products through nonbank channels without relinquishing ownership of risk, data, or customer relationships. Teams emphasized compliance‑by‑design principles, faster time to market, and lower integration costs, with pragmatic, revenue‑driven use cases spanning travel, point‑of‑sale finance, and SME lending. To fully capture the opportunity, firms must prioritize embedded finance use cases based on value and execution complexity and invest in ecosystem capabilities that support scalable distribution.
What struck me most this year was the depth and quality of the demos on show, and the sense that the ecosystem has moved into a more mature phase. I saw the foundations of a new wave of innovation — more pragmatic, more operational, and more aligned with real bank constraints. If you’d like to explore what these shifts mean for your organization, I’d welcome a conversation in an inquiry or guidance session.
