Instructional Design Review Assistant

Quick Start · Review, Challenge, Improve and Learn

A second perspective on your storyboard. The design decisions stay yours.

Install in your AI tool

Use the skill with Claude, ChatGPT or Gemini where custom Skills are available in your account. Availability and the upload controls can differ by account.

  1. Download the skill ZIP from your download page. This Quick Start is a separate guide. Keep SKILL.md and all three reference files together.
  2. Choose your setup below. Upload the ZIP or the extracted skill folder as supported by your tool.
  3. Enable or save the skill. Its installed name is storyboard-review-coach, version 1.2.2.
  4. Start a chat with a small storyboard excerpt. Add the audience, the intended workplace task and relevant constraints. Remove personal or sensitive project data first.

Claude

Open the custom Skills settings, upload the ZIP and enable the skill. Your account or organization needs to allow custom Skills. Claude's setup instructions.

ChatGPT

Open Skills from your profile menu, select New Skill and choose Upload from your computer. Upload the skill and start a new chat. If the options differ, check the features available in your account. OpenAI's skill installation instructions.

Gemini

In the Gemini web app, open Settings, then Skills and Upload. Select the ZIP or the extracted folder containing SKILL.md and the references. Review it, then choose Create. Gemini's setup instructions and availability.

Use a DOCX, readable PDF, screenshots or pasted text supported by your chosen tool. Ask it to identify any parts it could not read. The setup instructions do not represent a comparative test of the skill across AI tools.

Four ways to begin

Review

Use the storyboard-review-coach skill to review this storyboard excerpt. The learners are [audience]. At work, they need to [task]. Check whether the activities and feedback support that task. Point to specific screens, distinguish design issues from questions to verify, and tell me what already works. Don't rewrite anything yet.

Challenge

I want to challenge your recommendation for [screen or activity]. My reason is [context or constraint]. Reconsider the finding, suggest a different approach and compare the trade-offs. Tell me if the original approach is reasonable in this context.

Improve

I approve [specific direction]. Revise only [screen or activity], with a clear before-and-after comparison. Preserve confirmed facts and required terminology. Flag missing SME or policy details and effects on other screens, but don't change those screens.

Learn

From the design decision we just agreed on, suggest one question I can use in a future project. Explain when it is useful and when it may not apply. Keep it specific to the reasoning we confirmed, not a universal rule.

The assistant does not automatically remember an insight across projects. Save the question yourself if you want to use it again.

A full before/after example

Illustrative redesign, not a client result or a comparison test against ordinary Claude. The example decisions and feedback below require local SME/policy review.

The intended workplace performance

At a shared workstation, recognise that another person's session is active and decide what to do before starting your own task. Confirm the required action and any exceptions with your SME.

Before: explain the risk after the action

A worker uses a colleague's account to finish a task.

Learner task: List the risks of using someone else's account.

This asks learners to explain a risk after the decision has already been made. That can be useful when explaining risks is the objective. It does not let you observe what they would choose before acting at the workstation.

After: make the decision before acting

You arrive at a shared workstation to complete a task. Your colleague has left, but their session is still open. You can see their name in the corner of the screen.

Learner task: What would you do next, and which detail led you to that decision?

Start with a short written response, or use choices if they match your audience and assessment needs. Candidate choices to validate with the SME:

Feedback that supports the decision

Write feedback around the learner's action and the relevant policy. For example:

Do not publish these choices as scored answers until the SME has confirmed the rules, the plausible mistakes and the feedback. Do not invent a breach, disciplinary outcome or technical consequence.

What changed?

The learner now interprets a workplace cue and chooses an action before acting. The important change isn't the format. It's the thinking the learner has to do.

Challenge the redesign

For a novice audience, prior instruction or a worked demonstration may be appropriate. In another setting, a procedural job aid may be more useful than a decision scenario. Choose based on the task, the risks and the learners' starting knowledge.

A question to take forward

“Does this activity ask learners to make the decision they face at work, or only explain it afterwards?” Use it when the target performance involves a decision. It is not a reason to replace every recognition, recall or fixed-procedure activity with a scenario.

Your experience still matters

The assistant cannot verify your organization's rules, know whether a distractor reflects real workplace behavior, or prove that a design will improve performance. You'll still make those calls with your SME and project stakeholders where needed.

Want to take this further?

In the Learning Design Lab, you work on your own AI design process. Join the waitlist to receive the programme and enrolment details.

Join the Learning Design Lab waitlist

Mark Spermon · mark@markspermon.com
Quick Start updated 13 October 2026. Skill version 1.2.2, updated 13 October 2026 with programme and waitlist links.