CALM: Let the Interface Fail on Paper First

A calm interface often begins with the options the team chooses not to test.

A good heuristic review gives the interface a safe place to fail.

Heuristic evaluation compares an interface against a small set of usability principles before the team spends scarce test time on every possible variant. Under pressure, this reduces decision overhead by using relevant criteria, such as system status, consistency, error prevention, recognition, recovery, and workload.

Guardrail: Do not bring an interaction option into usability testing until it has survived a paper review against the tasks, risks, and heuristics that matter most.

In a patient transport scenario, acknowledging alarms must be quick without becoming casual. In one project, a team explored several touch interaction options for a new alarm interface: a direct button, a gesture, a press-and-hold, a swipe, and a compact confirmation pattern. Instead of prototyping every option deeply, the team built quick representations and scored them against criteria such as clarity, touch effort, recognizability, accidental activation risk, and expected error rate.

The weaker ideas did not need a full usability session to reveal their cost. They already created hidden checks, extra steps, or ambiguous system states on paper. The result was a smaller set of candidates that deserved real participant time, which made the next usability round sharper and easier to interpret.

In a logistics dashboard, a warehouse scanner, or an incident-management tool, the same move applies: evaluate the interaction pattern before the workflow absorbs it.

Every interaction variant that survives too long creates extra prototype states, test scripts, implementation branches, regression cases, and review effort.

Ship cue: log rejected interaction options in a lightweight decision record, including the violated heuristic, the risk it posed, and the reason it did not move into user testing.

Heuristic Triage Card
Before an interaction option reaches usability testing, score it on:

Nielsen Norman Group’s heuristic evaluation guide pairs well with this card because it keeps the review structured and prevents the table from becoming a debate about taste.

If less interface often means less attention cost, which interaction in your product exists mainly because the team liked it, not because the user needed it?

Today, let one idea fail quietly before you turn it into action.

What feels attractive about it?

What hidden effort would come with it?

Next time, notice that not every good idea needs to become part of your life.

AI can support heuristic evaluation as an early review layer, especially when teams compare interaction options before investing in prototypes or usability sessions. It can surface inconsistencies, missing feedback, unclear labels, weak error prevention, and avoidable workload.

A 2026 study on the Radiation Planning Assistant, a web-based radiotherapy planning tool, found that ChatGPT identified 26 potential usability issues, of which 9 were considered actionable by the developer team. The important limit: none were critical, and all actionable findings were minor. The practical lesson is calm and useful: AI can enrich early heuristic review, but it should not replace expert judgment, clinical context, or user testing.

For this episode’s principle, AI fits best before the team review. Give it the task, user role, screen flow, alarm context, and chosen heuristics. Ask it to find likely slips, violated heuristics, and questions for the team. Then let the humans decide which options fail on paper, which need a prototype, and which deserve real usability testing.

Best,
Andreas Walden

Share this with someone designing for clarity.