U.S. banks are adding conversational AI, stronger authentication and accessibility features to mobile channels, but adoption remains uneven, according to Keynova Group’s Q3 2026 Mobile Banker Scorecard. The study of 17 leading U.S. retail banks found that only Bank of America and U.S. Bank currently offer native voice interactions in their mobile banking apps, while passkey adoption remains significantly higher on mobile web than in banking apps.

Mobile banking is entering a new phase in which the user interface is becoming increasingly conversational, but Keynova Group’s latest benchmarking suggests banks are still moving cautiously from traditional app navigation toward voice-enabled and AI-assisted financial services.

The Q3 2026 Mobile Banker Scorecard, released September 22, evaluates more than 350 criteria across the mobile apps and websites of 17 major U.S. retail banks. U.S. Bank ranked first overall and also topped the mobile app and mobile web categories in the assessment.

The more significant finding for the wider banking technology market is the limited availability of native voice banking. Keynova found that only Bank of America’s Erica and U.S. Bank’s Smart Assistant currently support voice interactions inside their mobile applications for tasks such as transfers, bill payments and card controls. Bank of America’s Erica can additionally surface Zelle transactions and initiate certain peer-to-peer payments through voice interactions.

The gap is notable because conversational AI has become a familiar interface outside financial services. Keynova reports that 35% of the banks in its Scorecard now provide advanced virtual assistants using natural-language processing through text, but voice functionality remains much less common.

Both leading examples demonstrate that voice banking does not necessarily require customers to leave the authenticated banking environment. U.S. Bank’s Smart Assistant lets customers speak or type requests ranging from checking balances and credit scores to paying bills and transferring funds.

Bank of America similarly allows authenticated mobile users to interact with Erica through voice or text. The bank says Erica can help with account information, transactions and card management, while its service agreement notes that voice interactions are recorded and transcribed and that certain interactions can be retained for service improvement.

That security and privacy dimension is increasingly important as AI banking assistants move from information retrieval toward transaction execution. A chatbot that answers a question carries a different risk profile from an assistant capable of transferring money, paying a bill or changing a card setting.

Keynova’s findings show that authentication is evolving alongside conversational banking. Passkeys are now available for mobile web logins at 30% of Scorecard banks, compared with only 18% for mobile app logins. Nearly half of the banks also require multifactor authentication when customers reset a PIN, while only two require MFA to unlock a debit card, according to the research.

Citi is taking another route for some high-risk transactions, with the Scorecard identifying passwordless biometric or six-digit PIN authentication through its mobile application for wires and ACH transfers.

The move toward passkeys is also visible outside Keynova’s benchmark. U.S. Bank says its passkeys use the same device-level verification customers use to unlock a device, such as a face, fingerprint or device PIN, although its current rollout is for online banking rather than the mobile app.

The convergence of conversational interfaces and stronger authentication reflects a broader shift in banking technology. McKinsey reported in its 2025 Global Banking Annual Review that more than half of consumers surveyed were already using generative AI tools, while many expressed interest in AI capabilities from their banks. The research also identified mobile as the most widely used banking channel.

McKinsey’s more recent research on agentic AI found that 23% of surveyed consumers were already using generative AI for financial tasks at least monthly, while 57% said they would consider using a third-party generative AI financial agent if their bank did not provide one.

Those findings help explain why banks are experimenting with conversational interfaces even as governance requirements remain substantial. An AI assistant increasingly needs to understand intent, retrieve financial information, navigate banking journeys and potentially execute actions while maintaining authentication and audit controls.

Accessibility is another part of the equation. Keynova found that approximately half of the banks assessed offer mobile viewing adjustments such as text-size controls, dark/light modes and landscape display. Yet fewer than half provide direct links to accessibility policies on app and mobile-web login screens. Privacy and security policy access is also inconsistent between mobile apps and websites.

For banks, that creates a three-part mobile technology challenge: AI assistance, security and accessibility need to evolve together.

The industry is already moving in that direction. McKinsey’s research on mobile banking describes an emerging model in which conversational AI can execute transactions or route customers into the appropriate banking journey, while voice interaction becomes another interface for AI agents.

The commercial implications extend beyond customer experience. Better conversational interfaces could reduce navigation friction and shift routine service activity away from traditional call centers. McKinsey estimates that AI-enabled customer-care transformations can create substantial cost-reduction opportunities, although outcomes depend heavily on process redesign, data quality and governance rather than AI deployment alone.

The competitive issue is therefore no longer simply whether a bank has a mobile application. The emerging question is how effectively that application can understand customer intent, authenticate sensitive actions and provide assistance without sacrificing transparency or accessibility.

Keynova’s Q3 assessment suggests that U.S. banks are at different stages of that transition. Voice-enabled banking remains concentrated among a small number of institutions, while text-based conversational assistants, passkeys and layered authentication are spreading more broadly.

For digital banking providers, the next phase will likely involve connecting these capabilities rather than treating them as separate features. A voice request may need to trigger an AI reasoning layer, authenticate the customer, execute a transaction and provide an auditable confirmation—all within the same mobile experience.

That makes the mobile banking app increasingly less like a collection of screens and more like an AI-enabled financial services interface.

Market Landscape

Keynova’s findings arrive as banks face rising expectations for conversational and mobile-first financial services. Gartner has identified generative AI as an emerging opportunity for retail banking customer experience, while its 2026 research notes that the quality of a bank’s web or mobile app was the leading reason customers switched primary providers in 2025.

McKinsey likewise describes mobile as the primary banking channel for many consumers and identifies conversational AI, predictive engagement and voice-based agents as emerging components of mobile banking.

At the same time, security requirements make financial services different from general-purpose AI applications. Banks need to establish clear boundaries between conversational assistance and transaction authorization, particularly as assistants gain the ability to initiate payments, transfers and account changes.

The result is a banking technology market increasingly centered on AI orchestration, authentication, mobile experience platforms, fraud prevention and accessibility rather than on individual app features alone.

Top Insights

• Keynova found that only Bank of America and U.S. Bank currently offer native voice-enabled banking assistants within their mobile applications.
• Advanced text-based natural-language assistants are more common, with 35% of Scorecard banks offering this capability.
• Passkeys remain more established on mobile web, available at 30% of Scorecard banks versus 18% for mobile apps.
• Approximately half of the assessed banks offer mobile accessibility adjustments, but policy transparency remains inconsistent across channels.
• McKinsey research indicates growing consumer use of generative AI for financial tasks, increasing pressure on banks to develop conversational interfaces.

Get in touch with our fintech expert