Despite accessibility standards, ATMs, branch and digital banking services can all fall short when it comes to meeting the needs of disabled people, especially when transactions cross from one channel to another.

October 14, 2026 | Joshua W.J. Brown, systems researcher
A bank can have an accessible app, ATM and a customer-service line and still fail the customer at the exact moment the transaction needs to cross from one channel to another.
That is the operational problem accessibility programs often miss. They measure the pieces separately: the mobile app, the ATM interface, the call center script, the branch experience. But a customer does not experience banking as separate audited objects. A customer experiences a chain. If authentication fails in the app, if the ATM depends on a mobile control that is not usable, or if the fallback requires a branch visit that destroys the independence promised by self-service, the chain has failed even if each channel can point to its own checklist.
The most important design question is therefore not simply, "Is this channel accessible?" It is, "When this channel fails for accessibility reasons, what equivalent path preserves the customer's transaction, privacy and independence?"
This is not a theoretical edge case. A 2026 Journal of Medical Internet Research scoping review of digital financial applications for people with visual impairment mapped 23 studies and found authentication-related barriers in 18 of them. Screen-reader incompatibility appeared in 17. The review identified visual CAPTCHAs, one-time-password workflows that require app switching, inaccessible biometric checks, session timeouts and unclear error messaging as recurring barriers.
For ATM and omnichannel operators, that finding should change the definition of uptime. If the machine is powered, networked and physically available, but the authentication path depends on a mobile screen, visual challenge, inaccessible code flow or unusable handoff, the service is not truly available to the affected customer. The breakdown is not only digital. It becomes a self-service failure, a privacy failure and often a dignity failure.
A customer who must disclose a code to a stranger, ask a family member to complete a transfer, or abandon the transaction and visit a branch has not received an equivalent banking service. The institution may still log the ATM as operational. The customer has experienced an outage.
The fallback path cannot be an apology after failure. It has to be part of the transaction architecture.
If a cardless ATM flow requires mobile-app confirmation, there must be an accessible alternative that can complete the same transaction with equivalent security. If a one-time password fails because the user cannot switch apps without losing screen-reader position or timing out, the system needs a nonvisual, time-tolerant method rather than another reset loop. If a biometric prompt cannot be completed by the user, the alternative should not downgrade them to a high-friction telephone process that cannot reproduce the transaction state.
The standard should be simple: every material authentication path needs an equivalent failover path. Equivalent does not mean identical. It means the customer can complete the same banking objective with comparable security, privacy, cost, timing and autonomy.
Banks already measure uptime, failed transactions, fraud rates, abandonment and service recovery. Accessibility-related authentication failures should be measured with the same seriousness.
A useful exception log would capture the channel, the failed step, the failure class, the fallback offered, time to recovery and outcome. Was the customer able to complete the transaction independently? Did the bank preserve the transaction state when the customer moved from app to ATM, ATM to phone, or phone to branch? Was the customer forced to disclose sensitive information to another person? Did a security control protect the bank by excluding the customer?
Those questions are operational, not sentimental. They tell a bank where the handoff is failing and which fixes actually preserve service.
Branches still matter, and human service can be a crucial accommodation. But a branch visit should not be the default fallback for every inaccessible digital or ATM failure. That converts a self-service accessibility defect into lost time, transportation burden and dependence on business hours.
The better model is layered recovery. First, make the primary path accessible. Second, make the equivalent fallback accessible. Third, provide a human escalation path that can see and preserve the transaction state instead of forcing the customer to start over. A staff member should not have to reconstruct the problem from scratch, and the customer should not have to narrate an accessibility failure repeatedly just to prove that the failure happened.
For ATM and omnichannel teams, the next step is not another general declaration that accessibility matters. It is a channel-handoff audit.
Pick a transaction: cash withdrawal, balance inquiry, transfer, cardless cash, PIN reset, failed login recovery. Map every authentication step. Then ask what happens if the customer cannot see the prompt, cannot solve the CAPTCHA, cannot switch apps fast enough, cannot use the biometric flow, cannot read the error message or cannot complete a timed code. If the answer is "call support" or "go to a branch," the design is incomplete unless that path preserves the transaction state and can complete the same banking objective.
Accessible banking is not a channel checklist. It is a continuity problem. The customer should not fall out of the system because the bank's app, ATM and support process were designed as separate kingdoms. The fallback path is part of the product. If it is missing, the service is not fully built.