Redesigning Test Case Planning with AI Assistance (2025)

Redesigning Test Case Planning with AI Assistance (2025)

Client
Confidential (Developer Tools Domain)
Timeline
6 Weeks
Role
UX Designer (research, concepts, flows, prototypes; partnered with engineers and PMs)
Tools
Figma, FigJam

Building test cases used to mean jumping across tabs, copying and pasting, and retyping the same details twice. I redesigned the workflow with a context-aware AI assistant, transforming it from 6 fragmented manual steps into 4 streamlined AI-assisted steps, giving developers speed, focus, and control.

6 manual steps 4 AI-assisted
| 0% faster (projected based on workflow step reduction)

‍

Disclaimer 📌

All visuals and product details are fully anonymized and recreated to respect confidentiality while accurately reflecting the challenges, design process, and impact of my work.

Problem 🧩

Test case planning is repetitive, fragmented, and slow.

“Building test cases by hand feels like doing the same work twice.” — Developer pain point (from stakeholder conversation)

This fragmented process creates recurring issues:

  • Breaks focus through constant context switching
  • Increases errors from duplication and missing coverage
  • Slows productivity with manual drafting and reformatting
See how the 6-step workflow plays out, with pain points at each stage.
  1. Hunt through PRDs to extract acceptance criteria.
    ⚠️ Context switching
  2. Infer scenarios from criteria and personal heuristics.
    ⚠️ Guesswork
  3. Manually draft titles and descriptions.
    ⚠️ Manual drafting
  4. Specify environment and parameters.
    ⚠️ Setup overhead
  5. Self-review for coverage and duplicates.
    ⚠️ Rework
  6. Share for peer review.
    ⚠️ Peer latency

‍

Pain Point → Research Insight → Design Goal 📊

From stakeholder conversations and workflow analysis, key themes emerged. I aligned each pain point with supporting research insights and proposed design goals to turn them into actionable solutions.

‍

1
❗ Pain Point

Fragmented copy-paste across tools

Developers switch between PRDs, test plans, and trackers to move the same details.

🔎 Research Insight
“Most planning time is spent reformatting and pasting from other docs.”
Design Goal

AI in the flow of work

Surface AI help directly in context. No extra tabs.

2
❗ Pain Point

Review bottlenecks hide issues

One-by-one reviews slow releases and let problems slip through.

🔎 Research Insight
“Reviewing 100+ cases manually is unrealistic. We skim and miss things.”
Design Goal

Faster reviews

Batch actions and previews make large reviews quick and manageable.

3
❗ Pain Point

Speed vs. oversight trade-off

Teams want speed but also need visibility and trust in results.

🔎 Research Insight
“Automation helps, but I need to see what changed and why.”
Design Goal

Keep developers in control

All suggestions stay editable, reviewable, and rejectable. AI proposes, humans decide.

‍

Exploring AI Entry Points in the Workflow 🚪

I mapped potential moments where AI could assist during test creation through the lens of discoverability and focus.

The goal was to make assistance feel available yet invisible.

  • Global “Ask AI”

    Universal access for power users anywhere in the experience.

    • Always available
    • May compete with the primary task flow
  • Chat panel

    Expandable guidance and deeper exploration without blocking the workspace.

    • Persistent yet optional
    • Can be overlooked when collapsed
  • Inline icon

    Context-aware help near acceptance criteria, aligned to intent.

    • Quick and easy to find
    • Risk of visual noise if overused
Iterated concept sketch showing global entry, inline trigger near acceptance criteria, and a side chat panel.
Exploratory sketch of AI entry points in the test planning UI, comparing global, inline, and side panel access.

[change chat panel to side panel]

‍

Early Concepts 🧪

Concept 1: Global “Ask AI” entry point

  • ✅ Always accessible within the workflow
  • ❌ Poor context-awareness
  • ❌ Felt disconnected, required extra effort


Concept 2: Inline trigger via acceptance criteria

  • ✅ Contextual to the task
  • ❌ Entry point was difficult to discover
  • ❌ Created a fragmented experience (modal broke flow)
Hand-drawn exploration of global Ask AI and inline modal concepts with notes on context loss and fragmentation.
Fig. 1 — early exploration of entry-point trade-offs: visibility vs. focus.
Early Concepts · what we learned

Concept 1 — Global Ask AI: highly visible, but detached from the working surface; increased context switching.

  • Discoverability ✔︎ · Trust ✖︎ (felt “beside the work,” not “in it”)

Concept 2 — Inline trigger: aligned with intent and felt natural, but the entry was too subtle and the modal broke flow.

  • Trust ✔︎ · Discoverability ✖︎ · Flow ✖︎ (modal interruption)
Core tension: make help visible enough to find, yet quiet enough not to distract.

Fig. 2 — the **progressive assistance model** finds the sweet spot between discoverability and focus.

Iterated concept sketch showing inline icon, side panel with adaptive prompts, and unified experience.
Fig. 3 — iterated concept: inline icon for quick assist, expandable side panel for depth.
Final Flow · progressive assistance
1

Acceptance criteria viewed

System detects relevant context while the developer writes criteria.

2

Context-aware inline icon

Lightweight affordance appears at the right moment without stealing attention.

3

Side panel opens prefilled

Prompts and draft cases arrive scoped to the task; nothing to retype.

4

Review with control

Edit inline, regenerate, or batch-accept. AI proposes — the developer decides.

‍

Initial Concepts

‍

⚡
Core Insight Both concepts revealed core issues: difficulty discovering the entry point and fragmented user flow.

‍

Iterated Concepts 🪄

I merged the best of both worlds: guidance and control.

  1. Acceptance criteria viewed Developers review criteria while planning test cases.
  2. Context-aware icon appears inline A floating icon surfaces at the right moment without disrupting flow.
  3. Chat panel opens with tailored prompts Clicking the icon opens a side panel preloaded with relevant suggestions.
  4. Developer reviews and refines AI drafts test cases instantly. Developers can accept, edit inline, or regenerate, staying in full control.

‍

Iterated Concept

‍

‍

“Now I can generate and review test cases right where I’m working instead of copying across tabs.”

— Developer POV (expected)

‍

Optimizing for Speed 🏎️

Once I identified clear AI entry points, I worked on making review more efficient. Developers could accept multiple test cases at once using batch selection. Expandable previews let them quickly scan suggestions. When edits were needed, they could either make changes inline or ask the AI to adjust content.

(Recreated illustration of inline editing and batch selections)

‍

Impact ⚡

✅Efficiency:
  • Reduced the test planning workflow from 7 manual steps to 4 AI-assisted steps (~40% fewer steps)
  • Improved review speed with batch actions, quick previews, and inline edits
✅Trust:
  • Integrated AI into the way developers already work, so it felt natural and familiarl
  • Kept suggestions transparent and easy to understand
  • Ensured developers stayed in control, with clear ways to review and refine suggestions
✅Scalability:
  • Created flexible UX patterns ready to extend across other developer tools

‍

Key Takeaways ⭐

  • AI should feel like a natural extension of existing workflows, not a disruption
  • Provide multiple intuitive entry points so assistance feels accessible but non-intrusive
  • Speed matters: batch actions, inline edits, and quick previews reduce friction and keep users in flow
  • Transparency and clear control build trust, especially when AI suggestions influence decisions
  • Scalable UX patterns help future-proof AI-assisted experiences across different tools and contexts

‍

Final Thought 💭

Good AI design is not just about adding intelligence. It is about integrating it in ways that feel natural, keep people in control, and make their work easier without getting in the way.

‍

Let's make something cool.

Tell me a bit about what you are looking for and I'll be in touch.
Thanks for dropping a line! I'll be in touch shortly.
Oops! Something went wrong while submitting the form.
Back to Top ↑

© 2025 Kendra De'Anne