These prompts help you use AI at key stages of tree testing — writing unbiased task scenarios, analyzing path data for patterns, generating alternative category labels for low-performing sections, and producing stakeholder-ready reports.
Generate tree-testing task scenarios
I am conducting a tree test to evaluate a website's information architecture. Here is the tree structure:
[Paste the full tree hierarchy, indented to show levels]
I need to test findability for these content items:
1. [Content item 1] — located in [Category > Subcategory > Target]
2. [Content item 2] — located in [Category > Subcategory > Target]
3. [Content item 3] — located in [Category > Subcategory > Target]
[Add 5-10 items]
For each content item, write a tree-testing task scenario that:
1. Describes a realistic situation where a user would need this information.
2. Does NOT use the exact category label or any words from the label that would give away the answer.
3. Is concise enough that a participant can read it in under 10 seconds.
4. Avoids burying the key information in a long story.
Also flag any tasks where avoiding label words is especially difficult and suggest how to handle them.Analyze tree-testing results and identify problem areas
Here are the results of our tree test. Analyze them and identify the most critical navigation problems.
**Tree structure:**
[Paste the tree hierarchy]
**Task results:**
| Task | Correct location | Success rate | Direct success | First click correct | Avg time (sec) |
|------|-----------------|-------------|---------------|-------------------|---------------|
| [Task 1 description] | [Path] | [X%] | [X%] | [X%] | [X] |
| [Task 2 description] | [Path] | [X%] | [X%] | [X%] | [X] |
[Continue for all tasks]
**Path data for failed tasks (top wrong destinations):**
- Task [N]: [X%] went to [wrong category], [X%] went to [another wrong category]
[Add path data for tasks with <70% success]
Analyze this data and provide:
1. A prioritized list of problem categories, ranked by how many tasks they affect and the severity of the findability failures.
2. For each problem category: what is likely confusing users (labeling issue, grouping issue, or depth issue) and a specific recommendation.
3. Categories that performed well and should be preserved.
4. An overall assessment of the tree's health (what percentage of tasks met the 70% success threshold).Suggest alternative category labels
In our tree test, the following categories had low success rates. For each one, suggest 5-10 alternative label options.
**Category 1:** [Current label]
- Located under: [Parent category]
- Contains: [List key content items inside this category]
- Problem: [Describe what happened — e.g., "Users expected this content under [other category]" or "Users did not understand what this label means"]
- Nearby categories: [List sibling categories for context]
**Category 2:** [Current label]
[Same structure]
For each suggested label:
1. The proposed label.
2. Why it might work better (what mental model does it match, what ambiguity does it resolve).
3. Any risk (could it cause confusion with another existing category?).
Prioritize labels that are concrete and specific over abstract and generic. Prefer language that matches how users talk about the content, not how the organization talks about it internally.Draft a tree-testing report for stakeholders
I need to write a findings report for our tree test. Here is the raw data and my notes.
**Study overview:**
- Purpose: [Why we ran this tree test]
- Participants: [N participants, recruitment method, target audience]
- Number of tasks: [N]
- Tree description: [Brief description of what the tree represents]
**Results summary:**
- Overall success rate (averaged across all tasks): [X%]
- Tasks meeting 70% success threshold: [X of Y]
- Most problematic tasks: [List 2-3 tasks with lowest success]
**Key findings from path analysis:**
[Paste your notes on where users went wrong and any patterns you observed]
**Moderated session highlights (if applicable):**
[Paste key quotes or observations from participants]
Write a report with:
1. Executive summary (3-4 sentences).
2. Per-task results table with traffic-light color coding (green >80%, yellow 60-80%, red <60%).
3. Top 3-5 findings, each with: the problem, the evidence (success rates + path data), and a specific recommendation.
4. What worked well (categories with high success that should be preserved).
5. Recommended next steps (retest after changes, additional research needed).