Product visibility in ChatGPT: what 3D graphics and animation have to do with it
Product visibility in ChatGPTWhat 3D graphics and animation have to do with it
In short: An AI assistant looking for an answer mostly reads the text of a page, not the video itself. It will recommend your product if it finds sentences that answer the buyer's question directly. 3D graphics and animation help twice: they give you a reason to build a dedicated product page, and their script and captions supply ready-made sentences about how the device works.
In this article:
How does an AI assistant choose the products it recommends in an answer?
More and more buyers start looking for a supplier by asking an AI assistant: "which heat pump for an industrial hall", "who makes milling cutters for aluminium", "what is the difference between two types of heat exchanger". ChatGPT, Gemini and Google's AI summaries have no product catalogue of their own. They build the answer from pages they find on the web and judge to be reliable.
The good news: there is no secret trick here. The guidelines published by the makers of these tools say things a marketing team can check without a developer.
1. There is no separate "AI optimisation". According to Google Search Central, there are no additional requirements for appearing in AI summaries - the same principles apply as in regular search. Among them, Google lists making important content available in text form and supporting that text with good images and video.
2. The data describing the page must match what people see. The same source recommends that structured data (a description of the page for search engines, invisible to readers) matches the visible text.
3. ChatGPT has to be allowed in. As OpenAI's documentation states, sites that block the OAI-SearchBot crawler in robots.txt will not be shown in ChatGPT search answers.
Sources checked on 9 November 2026: Google Search Central, "AI features and your website"; OpenAI, "Overview of OpenAI Crawlers".
What does this mean for a manufacturer? An assistant will recommend your product only when your page contains text that answers the customer's question - what it is, how it works, how it differs from others and what it is for. A beautiful photo without a description is an empty space to it.
Why is a video on the page not enough for AI to understand the product?
A product video convinces a person in seconds, but to a search engine and an AI assistant it is mostly a file with a title. According to Google's video guidelines, a video needs its own page on which it is the main subject, and having that page indexed does not yet mean the video will be indexed. A blog post that only embeds a video does not replace such a page.
This is where 3D animation has an edge over camera footage: before it is made, someone has to break it down into scenes. The script is a ready list of what the product does - the flow, the assembly sequence, the layers, the movement of the mechanism. Move it onto the page as a description under the video, and the same work starts answering buyers' questions. That is how good technical animations work: the image shows, the text names what you see.
- 1. Mesh
- 2. Clay model
- 3. Render
The commercial for the KAAST brand (produced for Gdesz Film Studio) made a lathe the hero of the ad. Movement and proportions are settled at the mesh and clay stage - changes are cheapest then, because nothing has been rendered yet.
What do 3D graphics and animation add to a page that AI quotes?
AI assistants quote pages that answer a question specifically. 3D graphics help build those answers, because they show what customers ask about most often and what cannot be photographed. Give such an image a caption and a paragraph - those are what end up in the answer.
Four customer questions a 3D image answers
Examples from projects in our portfolio.
- What's inside?A cross-section or exploded view shows the layers and parts that disappear from view once assembled, and the caption names them. For Knauf Automotive, an exploded-view animation shows EPP components mounted inside a car bumper - a photo of the finished car cannot show them.
- Where does it fit and what does it replace?For Serwis-Kop, spare parts were visualised in the context of the machine rather than on a white background. You can see where they fit, what they replace and how they differ from a cheaper substitute - three questions customers also ask AI assistants.
- How does it work?Each scene of an animation is a step in how the device works. Described under the video, they become an answer to "how does it work" that an assistant can repeat word for word.
- Which variant should I choose?One model produces images of every version and finish, so each item in your range gets its own image and description, and customers compare them without guessing. That is how product visualisations for shops, catalogues and websites are made.
EPP components disappear under the bodywork once assembled. An image like this lets the supplier show customers where and why their product works - without taking the car apart.
Does your product's advantage not show in a photo?
Send us a CAD file or a few photos and tell us what your customers ask about. We will suggest which 3D image answers them best.
Get in touchbiuro@modelight.pl | +48 509 510 197
Which of our pages do people sent by ChatGPT land on?
We checked this on our own website. Between August and early November 2026, Google Analytics recorded 63 visits to modelight.pl from ChatGPT and 7 from Gemini. That is not much and we do not treat it as a statistic - but the pages those visits started on are telling.
Apart from the home page, they were mostly descriptions of specific projects - a series of animations for the KAAST brand, an air-spring animation for Tiptopol - and the industry pages for medical and HVAC. What they share is a specific product, shown with animation and described in plain text: what it is, what problem it solves, what you see in the frame. An AI assistant sends people to where it found an answer, not to where the prettiest video is.
Source: Google Analytics 4 for modelight.pl, traffic sources chatgpt.com and gemini.google.com, 1 Aug - 7 Nov 2026.
What it looks like across industries
Excerpts from animations and visualisations in our projects. Each shows something a camera cannot capture, with a sentence underneath that an AI assistant can quote.
Customers see what they pay for
A premium tool earns its price in the geometry of its edges and inserts. A 3D close-up shows this to a process engineer better than a catalogue photo and gives sales a real argument.
Animation for CNC →A procedure explained without a model in hand
A doctor explains the procedure to a patient with one image, and a distributor presents the whole implant system without a physical model.
Medical animation →A decision before installation
Customers see the device in an interior like their own and can judge whether it fits - before they order installation.
Visualisation for HVAC →Every part in its place
An exploded view replaces the assembly drawing in a catalogue or manual. Service technicians and customers know at once which part is which.
Technical animations →Replacement without calling service
Customers know what to replace and where before they open the housing - one image replaces a paragraph of the manual.
3D cross-sections →
A spare part in its place
For Serwis-Kop we show parts in the context of the machine. Buyers see where they fit, what they replace and how they differ from a cheaper substitute.
Serwis-Kop project →What should you check on your product page before asking ChatGPT about it?
The simplest test: ask an assistant the question your customer would ask and see whether it names your product. If it does not, the list below shows where the cause usually lies. Each row can be fixed without rebuilding the whole page.
| Page element | What an AI assistant "sees" | What to improve |
|---|---|---|
| Product video | mainly the title and surrounding text | describe the scenes under the video; give the key video its own page |
| Render or packshot | file name, alt text and caption | a caption with the product name and one feature visible in the image |
| Cross-section and exploded view | only what is named in text | names of layers and parts in the caption or in the paragraph next to it |
| Structured data | product and video description for search engines | consistent with what is visible on the page - no promises beyond the text |
| robots.txt file | whether OAI-SearchBot is allowed in | ask whoever runs your site to check it is not blocked |
If your product so far exists only as engineering documentation, there is no need to wait for a photo shoot. A model built from CAD files gives you images and animation before the first unit is made - we have described separately how a CAD file becomes a render. And when a product has hidden internals, a 3D cross-section with labelled layers gives you the most concrete material.
Start with one product - the one customers ask about most. One well-described product animation on its own page will do more than ten videos without a word of commentary.
Let's show your product so customers and AI assistants understand how it works
Tell us what the product is and where it needs to appear - we will propose a 3D image and a description that answers your customers' questions.
Ask about a projectbiuro@modelight.pl | +48 509 510 197
Frequently asked questions
Does ChatGPT watch the videos and images on my website?
In search it relies mainly on the text of the page. Video and images help when they are described: with a title, a caption, a paragraph next to them and structured data that matches what is visible.
Do I need separate optimisation for AI assistants?
According to Google Search Central there are no additional requirements for appearing in AI summaries - the same principles apply as in regular search. For ChatGPT it is also worth checking that robots.txt does not block the OAI-SearchBot crawler.
How does 3D animation help product visibility in ChatGPT?
An animation script breaks the way a product works into steps, and a cross-section or exploded view names parts you cannot see. Moved onto the page as a description, they become text an AI assistant can quote when answering a buyer's question.
Is it enough to upload the video to YouTube or Vimeo?
A video on a video platform has its own description, but it does not replace a product page. It works best when the video is embedded on a page dedicated to it, with text underneath describing what you see in each scene.
Where should I start if I have many products?
With one - the one customers ask about most, or the one hardest to show in a photo. Once its page has an image, a description and answers to typical questions, the same layout is easy to repeat for the next items.