Harish Pillai examines how design decisions in navigation apps, ATMs, and medical devices affect safety and user confidence, and proposes concrete practices for building interfaces that keep control in the hands of users while reducing error risk.
Control, Usability and Accessibility: A Deeper Look
{{IMAGE:2}}
By Indie designer – Harish Pillai
May 15 2026
The Waze tragedy that sparked the conversation
In early 2026 a senior couple set out to visit their daughter using a popular navigation app. The route displayed several streets that looked identical to the one they intended to take. The app failed to highlight the subtle differences, and the couple turned onto a dead‑end that led to a fatal accident. The incident was traced back to a UI flaw: the map view offered no clear visual cue that the selected street was a different segment from the intended one, and the turn‑by‑turn prompts lacked redundancy.
The tragedy illustrates a core principle that designers sometimes overlook: control belongs to the user, not to the system. When an interface hides uncertainty, it forces users to trust the software blindly. In safety‑critical contexts—whether a driver on a rural road or a clinician operating a ventilator—this trust can be costly.
Lessons from the early ATM experience
In the 1970s a major bank introduced the first automated teller machines. The machines were built with a hard‑line safety rule: if a user entered the wrong PIN three times, exceeded the withdrawal limit, or attempted a transaction that the account could not cover, the machine would retain the card. The intention was to protect the bank from fraud, but the outcome was a wave of angry customers who lost their only means of accessing cash and often closed their accounts.
The failure was not technical; it was a usability mistake. The design gave the bank absolute control over the user’s access, without any graceful fallback. When the bank later released a version that simply ejected the card after a failed attempt, user satisfaction rose dramatically and the brand even re‑positioned itself with the slogan “The Citi never sleeps.”
What changed?
- Visible feedback – the new machines displayed a clear error message and a countdown before ejecting the card.
- Error recovery – users could retry the PIN without losing the card, reducing frustration.
- Consistent mapping – the physical act of inserting the card matched the expectation of getting it back, reinforcing the mental model.
These three adjustments turned a punitive system into a cooperative one, and the lesson carries over to any interface where a mistake can lock a user out of a critical workflow.
Healthcare devices: where control meets life‑or‑death stakes
Medical equipment designers face the toughest version of the same problem. A clinician must interpret sensor data, adjust dosage, and confirm actions under time pressure. Any ambiguity can lead to medication errors or equipment misuse.
A concrete example: infusion pump UI
An infusion pump that delivers medication intravenously often presents a numeric keypad, a flow‑rate display, and a start/stop button. In a poorly designed version, the start button is placed next to the stop button, both sharing the same colour, and the flow‑rate field does not update in real time. A busy nurse may inadvertently press stop, thinking they are confirming the dosage, and the patient receives no medication.
How a better design restores control
- Spatial separation – place start and stop on opposite sides of the screen, using contrasting colours (green for start, red for stop).
- Live feedback – update the flow‑rate display instantly after each keypad entry, and show a countdown timer.
- Confirmation dialogs – require a brief “Are you sure?” prompt when the user attempts to stop an active infusion, with a clear visual of the remaining dose.
- Error‑prevention mode – lock the keypad after a certain number of invalid entries and display a clear instruction to call technical support.
These measures keep the clinician in the driver’s seat, making the system’s state transparent and recoverable.
A framework for designing control‑first interfaces
From the three domains above—navigation, banking, and healthcare—we can extract a reusable checklist:
| Principle | Description | Practical tip |
|---|---|---|
| Visible mapping | Every action should have an obvious, immediate visual response. | Use animations or colour changes that confirm a tap or click. |
| Error tolerance | Users should be able to recover from mistakes without severe penalties. | Provide “undo” or retry paths that do not lock the user out. |
| Consistent metaphors | The physical or digital metaphor must match user expectations. | Align button placement with real‑world analogues (e.g., a lever that moves up to start). |
| Redundant cues | Critical information should be presented in more than one way. | Combine icons, text, and haptic feedback for important actions. |
| Graceful degradation | If a feature fails, the system should still allow core tasks. | Fallback to a simplified view when network latency is high. |
Applying this checklist early in the design process can surface hidden risk areas before a product reaches users.
Why the industry still struggles
Even with these guidelines, many products launch with incomplete usability testing. Companies often adopt the mantra “don’t fix it if it ain’t broken,” assuming that a low number of reported incidents means the design is sound. The Waze incident shows that low‑frequency, high‑impact failures can slip through because they happen under rare conditions—identical street layouts, poor signal, or user fatigue.
A more disciplined approach involves scenario‑based testing: simulate edge cases, such as multiple similar streets, low‑visibility environments, or users with limited motor control. Collect quantitative data on error rates and qualitative feedback on mental models.
Closing thoughts
Control, usability, and accessibility are not separate silos; they are interdependent facets of a user’s experience. When an interface respects the user’s agency, it reduces error, builds trust, and ultimately saves lives. Designers who view their work as a partnership with users—rather than a one‑way instruction set—will create systems that endure beyond the next product cycle.
Harish Pillai writes about the intersection of design, technology, and human factors on his blog Use Your Experience.

Comments
Please log in or register to join the discussion