Express
ATM Interface Acceleration
Primary embodiment from the patent — individual ATMs learn per-card withdrawal habits and skip to the completion screen after authentication.
A cognitive platform that learns your navigation habits across a software interface, predicts your destination screen, and takes you there directly — skipping every intermediate step automatically, with no manual shortcuts or configuration required.
Enterprise and consumer interfaces force users through identical multi-step navigation sequences on every visit. The same welcome screen, the same main menu, the same sub-menu — regardless of how many times you've done this before and how predictable your destination is.
The cognitive automation platform sits between the user and any software interface. It records the screens a user visits, builds an individual model of their navigation habits, and — once confident in the prediction — generates a customized interface that takes them directly to their destination without any intermediate stops.
The platform is not a replacement for the underlying GUI — it's an intelligence layer that observes interactions and generates optimized interfaces on top. The original screens still exist; the platform simply determines when and whether to skip them.
Each time a user completes a navigation sequence, the platform adds a record — screens visited, in order, final destination, time of day, session context. After enough sessions, patterns emerge: this user always ends at Screen F after touching Screen A and B, and never visits C, D, or E en route.
The model is per-user and continuously updated. A user who changes their behavior will see the platform adapt — old patterns fade, new ones accumulate confidence.
In the ATM use case: a user who always withdraws $200 at this terminal normally navigates Welcome → PIN → Main Menu → Withdraw → Amount → Confirm → Complete. The cognitive platform learns this and renders Welcome → PIN → Complete — skipping the four intermediate decision screens entirely.
After presenting an accelerated interface, the platform asks for feedback. The user confirms or rejects the prediction. Confirmations increase future confidence; rejections add weight against that shortcut. This closed-loop reinforcement keeps the model accurate as users' behavior evolves over time.
Crucially, the platform does not require explicit user programming. Users never create rules — they simply interact, and the system learns.
The patent's primary embodiment is an ATM that learns individual withdrawal habits. A customer who always withdraws $200 on Tuesday mornings is taken directly from authentication to the confirmation screen — four screens disappear, and the transaction completes faster with less interaction.
The platform adapts per customer at the individual card level. The same ATM presents a different optimized flow to each cardholder — without any pre-registration or preference settings from the customer.
The cognitive platform is interface-agnostic — any software with multi-step navigation flows and identifiable users is a candidate. The more repetitive the navigation, the more value the platform delivers.
The cognitive navigation architecture has attracted citation from financial institutions and technology companies in adjacent AI-driven interface work. Citation data verified via Google Patents, June 2026.