You paid for customer films. Real customers, real driveways, real smiles. Now someone asks whether ChatGPT or Google's AI answers can use any of it.
Fair question. Here is a careful answer, and a plan that works whether or not the answer turns out to be yes.
Can AI tools actually watch your customer videos?
Plan on no.
Nobody outside the big AI companies knows exactly what an assistant reads when it names a local business. They do not publish the recipe, and it changes. Anyone who tells you they know precisely how ChatGPT treats your video file is guessing.
What we can say with confidence is what a machine reads most reliably: text on a page, facts that agree with each other, and markup that labels things clearly. A video file is none of those.
So the safe position is simple. Treat the video as something for people. Treat the words around it as something for everyone.
Why the words around a video matter more than the video
A film of a happy homeowner is persuasive to a human in about ten seconds. To a machine reading your page, it may be a box with a play button.
The page text is what's left. If the only thing near the film is a heading that says "Testimonials," you have given a search system almost nothing to quote. If the page says who the customer is, what the job was, and what they said, you have given it a sentence it can use.
This is also a question of effort. Text costs almost nothing to add. A new film costs real money. Get the cheap thing right first.
How to write a short summary under each testimonial
Put two or three sentences under every film. Plain language, no hype.
A good summary covers:
- Who the customer is, by first name and town if they agree to it.
- What the job was, in the words a neighbor would use.
- What they said, as close to their own phrasing as you can get.
If you can, add a short transcript in a collapsed section below. That helps anyone who watches with the sound off, which on a phone in a waiting room is most people.
Do not invent praise. Do not polish a quote into something the customer never said. Use what they actually said, because that is the point of a testimonial.
Why you should say where each review was first posted
When we rebuilt the site for A New Leaf Painting, the company's reviews from Google, Facebook, Houzz and Yelp came onto one page, sitting beside the customer films. Each review was marked with where it was first posted.
That small label does real work.
A reader sees that the praise did not originate on the contractor's own site, which is what makes it believable. A machine gets a clear, consistent statement of provenance. And you avoid the quiet problem of a review appearing to come from nowhere.
If you gather reviews from several places, do the same. Name the source on every one.
How one consistent review figure helps search markup agree
Here is a mistake we see often. The homepage says 4.9 stars. The contact page says 5.0. The markup behind the scenes says something else again.
Humans shrug. Machines see disagreement, and disagreement is a reason to trust a page less.
On the A New Leaf site, the review figure comes from one place. Every page shows it, and the search markup uses the same number, so they cannot drift apart. The same thinking applied to business facts generally, from the founding year to the insurance limit. One source, many pages, no contradictions.
You can do this without a rebuild. Pick the figure, write it down, and make every page and profile match it.
Where should videos live on your site?
On a page where the surrounding text can do its job.
Our advice, in order of importance:
- Give testimonials their own page, with the film, the summary and the review sources together.
- Repeat one or two on the pages that sell, such as a service page, with their summaries attached.
- Keep a short one near your estimate button, because that is where doubt is highest.
Avoid burying films in a slider that loads after the page does. Text hidden behind a script is text some tools never see.
If you run a contracting business, our contractor website design work covers this layout in more detail.
What markup to add, and what to skip
Markup is the labeling layer. It tells a machine what a thing on the page is.
Worth adding:
- Business details that describe your company once, correctly.
- Video markup for each film, with a real title, a written description and a thumbnail.
- Review figures that match what a visitor can see on the page.
Worth skipping:
- Review markup for praise that does not appear on the page. Mark up only what a visitor can read.
- Invented or rounded star figures. If the number is not true, it should not be in the code.
- Stacks of duplicate descriptions of the same business. Describe it once.
Do not expect markup to produce a particular result in a particular tool. Treat it as clear labeling, not a lever. For the wider technical picture, see our SEO and analytics work.
How to test whether your testimonials are being picked up
Skip the theory and run a test. It takes twenty minutes.
- Ask ChatGPT and Google's AI answers the questions your customers ask, such as who to call for a job like yours in Raleigh or Cary.
- Note whether your business appears, and what is said about it.
- Compare the wording to your page. Do your summaries or review sources show up in the phrasing?
- Repeat in a month, after the written summaries are live.
Answers vary from one run to the next, so look for patterns, not a single result. If your business never appears, the film is not the first thing to fix. Your reviews, your listings and your page text come first, and our reputation management service handles that groundwork.
What to do this week
Pick your three best customer films. Under each, write a two or three sentence summary, add the source of any matching review, and make sure the star figure matches everywhere it appears.
Then run the test above and save the results. You will know more in a month than any prediction can tell you today.