5 prompts · Explainer video
AI prompts for explainer and product-demo video
You need to make something complicated understandable in two to three minutes.
Explainers fail by starting with the mechanism instead of the problem, and by using an abstraction the viewer cannot picture. These prompts force a concrete opening scenario, one governing analogy carried through the whole video, and a plain-language pass that strips jargon.
SaaS product explainers and demosTechnical concepts for non-technical viewersLanding page and onboarding video
The prompts
1. Problem-first explainer script
Claude or ChatGPTThe main explainer structure
Write a [2/3]-minute explainer script for [PRODUCT OR CONCEPT], for [AUDIENCE] who currently [WHAT THEY DO INSTEAD].
Structure: a concrete 15-second scenario of the problem happening to one specific person → why the current workaround fails → what this does differently, in one sentence → how it works in three steps → what changes for them → one clear next step.
Rules: no feature lists, no jargon the audience would not use themselves, one governing analogy carried through the whole script, and every abstract claim followed by a concrete instance.
In practice: Write a 2-minute explainer script for an API status monitoring tool, for engineering managers who currently find out about outages from customer tickets.
2. Analogy generator
Claude or ChatGPTFinding the metaphor the whole video hangs on
I need to explain [CONCEPT] to [AUDIENCE] who already understand [THEIR DOMAIN].
Give me 8 analogies drawn from their domain. For each: the mapping, where it holds, and — importantly — where it breaks down and would mislead them. Then recommend the one that survives the most scrutiny.
In practice: The 'where it breaks' column is what stops a clever analogy from creating a support ticket later.
3. Screen-recording demo script
Claude or ChatGPTNarrating a product walkthrough
Write a screen-recording demo script for [PRODUCT]. The user's goal is [GOAL].
Format as a table: what is on screen, the exact click or action, the narration line, and the on-screen callout text.
Rules: narrate the intent before the click ("to catch this early, I'll set a threshold") not the mechanics ("now I click settings"). Never narrate anything the viewer can see. Maximum 8 steps — if the flow needs more, tell me which to cut.In practice: Intent-before-click is the difference between a demo and a tutorial nobody finishes.
4. Jargon strip
Claude or ChatGPTMaking the draft comprehensible
Rewrite this script at a level [AUDIENCE] can follow with no prior knowledge of [DOMAIN].
List every term you replaced and what you replaced it with. Do not simplify the actual claims — only the language. Flag anything that cannot be simplified without becoming wrong, and suggest how to introduce it properly instead.
Script: [PASTE SCRIPT]
In practice: Ask for the replacement list — sometimes the jargon was load-bearing.
5. Motion-graphic direction
Claude or ChatGPT, then Lumen5 or invideoSpecifying the visuals for an animated explainer
For each paragraph of this explainer script, describe the motion graphic: what appears, how it animates in, what the viewer's eye should follow, and what stays on screen as an anchor.
Rules: one new visual idea per paragraph, build on the previous frame rather than replacing it where possible, and no visual may require more than 4 elements.
Script: [PASTE SCRIPT]
In practice: Building on the previous frame is what makes an explainer feel like one argument rather than a slideshow.
Tools that finish the job
Where these prompts go next
A prompt produces the script, the shot list or the markup. These are the tools that turn that output into a finished video.