To explain an AI product in 60 seconds, show it doing one real job from start to finish. Skip the model, the robot and the glowing brain. Name the problem, say what the product does in plain words, show one real input turning into a useful output, say where the user stays in control, and end with one clear next step. That fits in about 150 words of voiceover.
The reason this works is trust. People are now more wary of AI claims than excited by them, so the job of a 60-second AI explainer is not to impress. It is to make the product feel useful, specific and safe.
Key numbers at a glance
Why is an AI product harder to explain than other software?
An AI product is harder to explain because viewers arrive skeptical and the product often has no obvious screen to show. You have to earn trust and make an invisible process visible, in the same minute.
The skepticism is measurable. In Pew Research Center’s June 2025 survey, 50% of US adults said they were more concerned than excited about AI, up from 37% when Pew first asked in 2021 (Pew Research Center, March 2026). Among US adults aged 18 to 34, that share rose from 40% in 2024 to 55% in 2026 (Pew Research Center, September 2026).
The second problem is that AI often works out of sight. A spreadsheet app has a grid. An AI agent that triages support tickets has, at best, a log. If the video only shows a chat box and a spinning loader, the viewer learns nothing about what changed for them.
What is happening in AI marketing right now?
Buyers in 2026 are researching AI products with AI, distrusting AI-made content, and asking for proof and human control. Each of those shifts changes what your explainer has to say in its first minute.
- Buyers start with a chatbot. In G2’s March 2026 survey of 1,076 B2B software buyers, 51% said they start research with an AI chatbot more often than with Google, and 69% chose a different vendor than they had planned because of a chatbot’s recommendation (G2, April 2026). G2’s Tim Sanders put it this way: “Buyers have moved from reference to inference.” Your video is often what they watch after the chatbot has shortlisted you, to check that the summary was true.
- AI-made content is losing trust. Gartner’s October 2025 survey of 1,539 US consumers found that 50% would rather buy from brands that do not use GenAI in consumer-facing messages, ads and content, and 68% frequently wonder whether what they see is real (Gartner, March 2026). In a separate March 2026 Gartner survey, 49% of US consumers said GenAI has made content quality worse (Gartner, June 2026).
- Proof beats promises. Forrester found that nearly one in five business buyers (19%) trusted their purchase decisions less after AI tools gave them wrong or misleading results, and it forecasts that over half of buyers will rely on product trials before buying. Its 2026 predictions say “B2B buying will be defined by tangible evidence, accountability, and continued buyer empowerment” (Forrester, reported by Digital Commerce 360, October 2025).
- People want a human in the loop. In Gartner’s February–March 2026 survey of 3,566 B2B and B2C customers, 87% said companies using GenAI for customer service must offer a way to reach a human (Gartner, August 2026).
Gartner survey of 1,539 US consumers, October 2025.
Source: Gartner, March 2026
Gartner’s Emily Weiss sums up the opportunity: “The brands that win will be the ones that use AI in ways customers can immediately recognize as helpful.” A 60-second explainer is one of the fastest ways to make that help visible.
What should a 60-second AI explainer say?
A 60-second AI explainer should say five things, in this order: the problem, what the product does, one real task shown on screen, where the user stays in control, and one next step. At about 150 words per minute, that is roughly 150 words.
That word budget comes from voice pace: average American English conversation runs at about 150 words per minute, according to the National Center for Voice and Speech (VirtualSpeech). Our guide to how many words fit in a 60-second script covers the maths in detail.
| Beat | Time | Words | What to say |
|---|---|---|---|
| 1. Problem | 0:00–0:10 | 20–25 | The job your buyer hates, in their words |
| 2. What it does | 0:10–0:20 | 20–25 | One plain sentence: what goes in, what comes out |
| 3. Show one task | 0:20–0:42 | 50–55 | One real input becoming one useful output, step by step |
| 4. You stay in control | 0:42–0:52 | 20–25 | Review, approve, limits, data, human handoff |
| 5. Call to action | 0:52–1:00 | 15–20 | One next step: try it, book a demo, join the waitlist |
- Problem 0:00–0:10 20–25 words
- What it does 0:10–0:20 20–25 words
- Show one task 0:20–0:42 50–55 words
- You stay in control 0:42–0:52 20–25 words
- Call to action 0:52–1:00 15–20 words
Beat 4 is the one most AI videos skip, and it is the one that answers the trust gap above. It does not need to be long. “You review every reply before it is sent” or “Your data stays in your workspace” is often enough.
How do you show AI working without robots and glowing brains?
Show the input and the output, not the intelligence. A messy inbox becomes five drafted replies. A 40-page contract becomes three flagged clauses. The viewer sees the before and after, and the AI is simply the arrow between them.
Google’s People + AI Guidebook gives the same advice to product teams: “Focus on the benefit, not the technology,” and “Communicate the product’s capabilities and limitations clearly to set expectations, and do this early” (Google PAIR). It also notes that “example-based explanations are useful” when the reasons behind an AI’s output are hard to explain. A single, concrete example is exactly what a short video can carry.
Be careful with human-like AI characters. Nielsen Norman Group reported a 2025 study by Colombatto, Birch and Fleming with 410 US participants: the smarter people thought ChatGPT was, the more they took its advice, while seeing it as emotional predicted less advice-taking. NN/g’s conclusion: “Simulating an emotional, human-like persona is counterproductive” (Nielsen Norman Group, September 2025). If you use a character, make it a guide who shows the work, not a robot with feelings. Our post on designing a brand mascot covers this in more detail.
These swaps work in most AI explainers:
| Instead of | Show |
|---|---|
| A glowing brain or neural network | The real input (a document, a ticket, a call) |
| A humanoid robot | A simple cursor or guide character doing the task |
| “AI-powered” in big type | The time saved or the result, in the user’s words |
| A chat box with a loading spinner | The finished output, highlighted step by step |
| “Works with everything” | The two or three tools your buyers actually use |
What words should you avoid in an AI explainer script?
Avoid words that claim intelligence without showing it: “revolutionary,” “magic,” “autonomous,” “replaces your team,” and “AI-powered” as an opening line. Replace each with what the user sees happen.
Nielsen Norman Group’s Kate Moran warned in 2024 that “implementing cool new technology for its own sake has rarely worked out well,” and that it is up to designers to make AI features “actually seem useful” (Nielsen Norman Group, November 2024). The same applies to scripts. Big claims also carry real risk: in August 2025 the US Federal Trade Commission sued Air AI over deceptive claims about business growth and earnings; the FTC said the company marketed “conversational AI” that “can replace human customer service representatives” (FTC). Say only what the product does today, and show it.
A worked example: a 60-second script for an AI support tool
Here is a sample script for an imaginary AI tool that drafts support replies. Swap in your own product name where it says [Product]. It is about 135 words, which runs just under a minute at a relaxed pace and leaves a few seconds for the opening visual and end screen. Paste your own script into the timer below to check yours.
Problem (0:00–0:10): Your support inbox fills up overnight. By nine in the morning, customers are waiting, and your team is typing the same answers again and again.
What it does (0:10–0:20): [Product] reads every new ticket and drafts a clear reply from your own help center, written in your brand’s tone.
Show one task (0:20–0:42): A customer asks how to change their billing date. [Product] finds the right help article, drafts a short answer with the exact steps, and adds the link. Your agent reads the draft, changes one line, and sends it in seconds. By the time they look up, the next ticket is already drafted and waiting.
You stay in control (0:42–0:52): Nothing goes out without your team’s approval. Tricky or upset customers go straight to a person, every time.
Call to action (0:52–1:00): Connect your help center in minutes and see your first drafts. Start your free trial today.
Try it
Script timer
Paste your script. We’ll estimate how long the voiceover runs and whether it fits.
Short videos remain the format people prefer for this: in Wyzowl’s 2026 survey, 63% of people said they would most like to learn about a product by watching a short video (Wyzowl). On LinkedIn, the company’s own guidance is to “establish your point in the first 5 seconds,” and it notes that “a large portion of LinkedIn members will watch your ad with the sound off,” so burn in captions (LinkedIn Marketing Solutions).
How we help
We make AI explainer videos for startups that need a hard-to-see product to feel simple and trustworthy. We start by finding the one task that sells the product, then write a timed script around it, including the “you stay in control” line most teams forget. In the storyboard we turn the invisible parts into visible input and output, using simplified, animated UI instead of raw screen recordings, so the viewer looks exactly where we want. You approve the script and storyboard before animation starts. We then cut the main video into short versions for LinkedIn, your homepage and sales emails. To check costs for your length and style, use our price calculator.
Your 60-second AI explainer checklist
- The first line names the buyer’s problem, not the technology
- One plain sentence says what goes in and what comes out
- One real task is shown from input to output
- The script says where the user reviews, approves or reaches a human
- No glowing brains, robots or “magic” claims
- Every claim is something the product does today
- About 150 words, read aloud at a relaxed pace
- Captions burned in for sound-off viewing
- Exactly one call to action
Sources
- Pew Research Center, “What the data says about Americans’ views of artificial intelligence” (Faverio and Kikuchi, March 2026; June 2025 survey)
- Pew Research Center, “Young adults around the world have mixed feelings about AI” (September 2026; 5,119 and 3,488 US adults, 2026)
- Gartner, “Gartner Marketing Survey Finds 50% of Consumers Prefer Brands That Avoid Using GenAI in Consumer-Facing Content” (March 2026; 1,539 US consumers)
- Gartner, “Gartner Survey Finds 49% of U.S. Consumers Say GenAI Has Made Content Quality Worse” (June 2026; 307 US consumers)
- Gartner, “Gartner Survey Finds 87% of Customers Say Companies Using GenAI for Customer Service Must Provide Access to a Human Agent” (August 2026; 3,566 customers)
- G2, “In the Answer Economy, Don’t Win the Click — Win the Answer” (April 2026; 1,076 B2B software buyers)
- Digital Commerce 360, “Forrester: B2B buyers now demand proof, not promises, about AI” (Mark Brohan, October 2025)
- Google PAIR, People + AI Guidebook, “Explainability + Trust” (undated)
- Nielsen Norman Group, “Prioritize Smarts over Sentience to Increase Trust with AI” (Evan Sunwall, September 2025)
- Nielsen Norman Group, “AI Features Must Solve Real User Problems” (Kate Moran, November 2024)
- US Federal Trade Commission, “FTC Sues to Stop Air AI from Using Deceptive Claims…” (August 2025)
- Wyzowl, State of Video Marketing 2026
- VirtualSpeech, “Average Speaking Rate and Words per Minute” (citing the National Center for Voice and Speech)
- LinkedIn Marketing Solutions, “Best practices for video ads”