FREE VELLORA TOOL · LOCAL ONLY

Usability task builder

Write a realistic goal, check for possible interface hints, and keep success criteria in separate researcher notes.

Define one task

Inputs stay in this tab. Use fictional test data. The wording check flags possible hints; it does not certify that a task is neutral.

Your working draft

LOCAL · EDITABLE
A STARTING POINT, NOT A FINDING

A realistic task with private observer notes.

Use the example or add your own context to create an editable result.

  1. Describe the situation and the outcome.
  2. Check wording for interface hints.
  3. Record completion, assistance, and setup failures separately.

Capture the evidence.

After the task, record observations, assistance and alternative explanations.

Open the evidence log →

HOW IT WORKS

Give a goal. Observe the route.

This local tool assembles your scenario and goal into a participant task, then creates separate observer notes. It flags supplied interface labels and common English procedural phrases. It does not use AI, rewrite your input, or inspect a prototype.

The example is fictional: a teammate needs an assigned action after a meeting. “Open the Actions tab and click Assign” reveals the path. “Make it clear that Alex owns the draft due Friday” describes the outcome while leaving navigation open.

Some label matches are necessary context. A task about exporting a document can legitimately use the word export. Review the meaning, not just the flag count. Synonyms, implied routes, and other languages may escape this simple check.

Keep the observer notes private.

  1. Prepare the same starting state for each attempt, using safe test data.
  2. Define success as something observable, including persistence when relevant.
  3. Record whether completion needed assistance.
  4. Mark a broken prototype or loading failure as not assessable.
  5. After the attempt, ask what the participant expected to happen.

Method references: NN/g on task scenarios and Maze on usability task wording. Checked October 11, 2026. This tool’s assembly and wording checks are Vellora’s implementation.

Turn observations into a decision.

Save each attempt in the evidence log. For an AI product, follow the onboarding study to distinguish interface friction from an unusable model output. A few qualitative sessions do not estimate a population success rate.