For 77% of US banking executives, integrating new technologies with existing systems and tools is a major challenge in modernizing contact center operations, according to Deloitte’s 2026 research on AI-assisted customer service in banking.

The finding highlights a broader issue behind the rapid adoption of AI in financial services. The question is no longer simply whether banks should automate customer service, but which interactions should be handled by AI, which require human oversight, and which still need to remain human-led.

Deloitte describes this as a spectrum based on the complexity of each interaction. Simple and non-urgent requests, such as bill payments or address updates, can increasingly move toward AI-led self-service. More complex interactions, including payment issues or loan applications, may be managed through an AI-led model with a human in the loop. At the other end of the spectrum, highly sensitive or complex interactions, such as fraud or loan-related issues, are expected to remain human-led, with AI supporting the customer service professional.

This distinction matters because customer expectations do not disappear when automation increases. For banks, the challenge is therefore not to automate as much as possible, but to create an operating model that can move customers smoothly between different levels of AI and human support.

From automation to orchestration

This is where Covisian’s approach comes in.

Through its AI Operation Center, Covisian enables different levels of AI and human involvement to coexist within the same operating environment. Interactions can be managed through AI Only, AI Supervised and Human Only modalities, allowing the level of automation to adapt to the complexity, context and requirements of each customer need.

Routine requests can be handled directly by AI, while more complex interactions can move into supervised models where human professionals oversee AI-supported conversations. When greater judgment, empathy or accountability is required, the interaction remains human-led, with AI supporting the agent through access to relevant information, context and operational guidance.

This approach also addresses one of the key risks of fragmented automation: losing continuity when customers move between channels, systems or service models. By connecting AI, human agents, customer history and operational workflows, banks can reduce the need for customers to repeat information and create more consistent handoffs across the journey.

“Customer service in financial services is not moving toward one single level of automation. Different interactions require different combinations of AI and human expertise. The ability to move seamlessly across those models is where we see the real opportunity for banks today,” says Bart Miller, CEO of Covisian USA.

As banks continue to invest in generative and agentic AI, the ability to orchestrate these different models will become increasingly important. Deloitte found that 70% of banking executives expect AI to shift human agents toward higher-value roles, reinforcing the idea that the future of customer service is not simply about replacing human interaction, but about using it where it matters most.

For financial institutions, the next phase of AI adoption will depend on how effectively they can decide when AI should lead, when people should lead, and how seamlessly the two can work together.