AI Design · Foundations

A two-part assessment for designers shipping AI products. First a short adaptive quiz across prompt-interface design, uncertainty surfacing, generative UI, agentic UX, onboarding, and safety UX. Then a real prompt-interface you design, graded against the Design Foundations rubric.

How it works

Two parts. About 10 hours of focused work, spread over a few days.

Part 1 · Quiz

18 adaptive questions

~30 min · in one sitting

Multiple-choice questions across AI design microskills: prompt-interface design, uncertainty surfacing, AI feedback patterns, streaming and perceived latency, safety UX, generative UI, agentic UX, onboarding, and capability surfacing.

Part 2 · Project

Design the surface

~8h · over a few days

Design a prompt-interface for a non-expert user. Submit a static prototype, an interaction flow, a 60-second walkthrough, and a 600-word narrative.

Score

Composite, owned by you

0 to 100 · valid 18 months

Quiz contributes 40%, project contributes 60%. The Design Foundations rubric scores all four dimensions. Published to your verified profile.

The project

Design a prompt-interface for a non-expert user

Design the prompt-interface for one of three feature briefs we provide (see the kit, or pick your own with a one-paragraph user-and-domain statement at the top of your narrative). The user is explicitly non-expert — assume they have never used a chat-style LLM tool.

Required deliverables

Static prototype (Figma, Framer, or live HTML) showing: empty state with three contextual prompt chips, scaffolded input, streaming/in-progress state, output with inline citations or a paired uncertainty signal, refusal-as-redirect for one out-of-scope action, and an undo/regenerate/edit-and-retry surface.
Interaction flow (≤ 6 frames) for the primary path, plus one agentic-UX frame showing how a long-running task surfaces progress and offers cancel.
Annotations on every frame explaining the design decision and at least one rejected alternative.
60-second walkthrough video.
≤ 600-word narrative articulating: who the user is, the trust principle the design rests on, the riskiest design call you made and why, and one calibration moment in the onboarding.

Out of scope

Production code. Static prototype is fine.
A full design system. One feature surface is enough.
Brand exploration. Use a neutral system; the rubric does not score brand.

What we look for

Refusal-as-redirect rather than flat ''I can''t help with that''.
Honest uncertainty — inline citations or qualitative bands, not unsupported confidence percentages.
Visible agentic surface for any long-running task — named steps + cancel.
Capability legibility — the user can tell which tools/skills are running without reading a manual.

How it's graded

One rubric — Design Foundations — applied at full weight. Four criteria: interface clarity (25%), uncertainty handling (25%), feedback and recovery (25%), design narrative (25%). Each criterion is scored 0–5 with a written rationale by the grader.

How we grade

One rubric. Four dimensions.

Your project is graded against the Design Foundations rubric. Each criterion is scored 0–5 with a written rationale, then weighted to a 0–100 project score.

Design · Foundations

Criteria & weights

Interface clarity25%
Uncertainty handling25%
Feedback and recovery25%
Design narrative25%

Anti-gaming

We measure thinking, not speed.

Every quiz answer has a 4-second minimum review time. Anything faster is recorded but doesn't affect your rating. Each question caps at 90 seconds. The whole quiz session has a 30-minute wall clock — once it expires, you finish what you have. Retakes for any track open after 14 days.