Maveric executive urges banks to shift CIO agenda toward AI-first operations

Anoop Melethil, head of marketing at Maveric Systems, has called on bank technology leaders to move their strategic focus from digital channel efficiency to AI-led decisioning across core operations. In an external opinion published by Finextra, he argued that the next phase of banking technology will affect onboarding, financial-crime controls, customer service, personalization and software release cycles.

Melethil said banks have spent years improving mobile apps, web processes and self-service journeys, with the aim of reducing friction for customers. That investment remains useful, he said, but it does not address a larger shift from systems that execute predefined workflows to systems that can interpret context and support decisions dynamically.

Onboarding and risk controls

Customer onboarding is one area where Melethil said the change is most visible. Banks must verify identities, meet regulatory obligations, detect fraud and manage risk while customers expect fast digital services. In his view, many digital-first processes still leave customers to enter data, upload documents, fix errors and wait for checks to be completed.

An AI-first model, as described by Melethil, would use orchestration across existing records, registries, third-party sources, digital signals and internal systems. The aim would be to ask customers for less information when a bank can responsibly verify or infer it, while sending exceptions to compliance teams with relevant context.

He also linked AI adoption to know-your-customer, anti-money-laundering and financial-crime management. Melethil said these functions have often been run through separate systems, data stores and workflows, increasing customer friction and operational complexity. He argued that AI can help banks interpret risk signals across different data types and processes, with potential benefits including faster onboarding, fewer duplicate checks, fewer false positives, improved fraud-loss outcomes and lower long-term integration costs.

Service, personalization and software delivery

Melethil said bank customer support should also move beyond rule-based chatbots, which he said work only when a customer’s request fits a preset decision tree. He described AI-first engagement as multimodal, spanning voice, text, image and video, with the system able to detect confusion, change the form of explanation and give human agents full context when escalation is needed.

For personalization, Melethil said banks have often relied on campaign-led models and traditional next-best-action tools built around fixed assumptions. He argued that AI can identify life-stage and need-based signals across salary deposits, rent changes, product use, balance movements, digital engagement and other contextual data. He said personalization must remain explainable and aligned with customer need, rather than feeling intrusive.

Melethil extended the argument to banks’ internal technology organizations. He said AI can be used across the software development lifecycle, including requirements gathering from customer feedback, call centre transcripts, market inputs and app reviews. He also cited automated generation and prioritization of user stories, coordinated development and testing, documentation, integration mapping and agentic quality engineering that monitors code commits, maps changes to testing coverage and repairs scripts.

He cautioned that faster delivery in banking still requires quality, security, compliance and explainability. Melethil framed the CIO agenda around three priorities: build data foundations for dynamic decisions, require engagement across multiple modes rather than separate channels, and automate the full software development lifecycle rather than coding alone.

This story draws on original reporting from Finextra Research.