This post is inspired by this thread, titled “AI Video Marketing: Is This the Future for Dealership Sales and Service?” I came across it by chance, and it made me want to explore this idea further, pushing current AI models to see how far they can go.
I shared several of my early tests in that thread. While not perfect, they already look promising, so I decided to write a more detailed post on how to create these kinds of videos if anyone else is interested.
I truly believe that, if used correctly, these tools can help dealerships increase their sales. Videos are simply more engaging and leave a stronger impression on people browsing cars online. Because of that, I think they offer a clear advantage and are very likely to become popular soon.
Right now, the number of videos in car listings is still low mainly because the process is slow, tedious, and often produces poor results. If there was an easier way to make them, they would quickly spread everywhere. This post focuses on creating videos easily, with tools that are available to anyone and that are either free or inexpensive.
To make these videos, all you need are good-quality pictures. I cannot stress this enough: the better your source photos, the better your final video will look.
If you take poor photos, compress them, and then upload them again, the quality will degrade at every step. But if you start with clean, high-resolution images and keep them that way, your viewers will have a hard time telling whether it’s real or AI. The goal is for them not to notice any AI at all.
The great thing is that taking pictures is easy. Anyone can take a decent shot. That’s not the case with video. Walkarounds of cars often look shaky or unprofessional, and they can actually take away from the car instead of adding value. Either do a great walkaround or skip it completely.
By using AI, we can now create perfect walkarounds and cinematic movements that would normally require expensive equipment and a full setup. That alone is a huge benefit. With these AI models, we can take our high-quality pictures and turn them into professional-looking clips that truly showcase the vehicle.
I believe the best car sales videos include a host — usually a beautiful girl who presents the car and grabs attention from people scrolling fast. “Who is that beautiful girl next to the Ferrari?” That’s exactly the type of reaction we want.
The good news is that you don’t need to hire a model or a real person. It’s all AI.
The first thing we’ll do is generate our model. For this, you can simply use ChatGPT. Ask ChatGPT to generate the picture of the model however you want her to look. You decide her clothing, hairstyle, and overall appearance. Make it detailed. You can reuse the same model for multiple cars — she’ll become your own virtual influencer, always ready to present your next car.
Now that you have your female model and great photos of your car, it’s time to generate the first keyframe. Basically, we’re going to combine the model with one of your car pictures.
To do this, we’ll use an AI model called Nano Banana. It allows you to use multiple inputs (for example, your AI model and a car image) and create a single, combined output.
So here’s the idea: give Nano Banana the female model you generated with ChatGPT and one of your car pictures, and then add a prompt like:
It might take a few tries to get the right image, but once you do, you’re ready for the next step.
Next, we’ll use an AI model that turns an image into a video. There are several available, and you can find the best-performing ones on this list:
Image-to-Video Arena | LMArena
For my tests, I’m using Veo 3.1, as it’s currently giving the best results. However, this changes fast, so if you’re reading this later, check that leaderboard for newer models.
Once you’re set up, generate your first clip. You’ll usually want to make multiple clips of the same scene to add variety. You can also generate clips without the female model, for example, to show the car’s interior.
At this point, it’s important to understand prompting — how to describe what you want the AI to create. The clearer and more specific your description, the better the results.
Prompting is basically giving instructions to the AI so it knows exactly what you want. A vague prompt will give you random results, while a detailed one will give you professional-looking output.
A bad prompt example:
A good prompt example:
A better prompt if you want to include your AI model:
Tips for writing good prompts:
Once you learn to prompt properly, the quality jump is dramatic. The same AI model can go from average to amazing just by improving your prompt.
Next, you need a script — a short text that will be turned into voice. You can use ChatGPT for this part too. Give it some details about the car (make, model, year, key features) and ask it to write a friendly, conversational sales script.
Once you have the script, use ElevenLabs to generate the audio. It will convert your text into a realistic, natural-sounding voice.
Then, overlay the voice track on top of your clips. You can use CapCut to edit them online, or any other editor you prefer.
At this point, you’ll have a nice-looking video, with different clips and your virtual model talking about the car. But there’s one problem: the lips don’t match the narration.
To fix that, use a lip-sync AI tool like Hailuo AI LipSync. Just search for it online. It’s simple to use — upload your video and audio, and it will synchronize the lips automatically.
How long this takes depends on how many clips you create and how polished you want your final video to be. Here’s a realistic estimate:
So a complete video, with multiple clips and proper editing, usually takes 2.5 to 3 hours total.
As for cost, it depends on the models you use:
In total, expect about $20 in AI costs per video. If you add an additional (you also need time to generate the clips with ai) 1.5 hours of editing time, a finished video usually costs around $60–100, depending on who edits it and local rates. The time and costs to edit are the main costs., not the ai models.
That means for under $100 and around three hours of work, you can produce a professional-quality sales video that looks like it came from a full production team.
Much like detailing, video content becomes more impactful as the value of the vehicle increases. Buyers spending $20,000 or more expect polished presentation. A professionally crafted video, even if AI-generated, signals effort and care — the same way a freshly detailed car does.
Here’s a quick guide:
A traditional car sales video produced by a local videographer can cost anywhere from $500 to $1,500, depending on quality and editing. That’s perfectly justifiable for a $60,000 exotic — but probably overkill for a $15,000 sedan.
AI bridges that gap. For under $100, you can create a virtual host, insert them into polished clips, sync the voiceover, and produce a clean, modern presentation — something that normally requires a videographer, model, gear, and studio.
Just like a $200 detail can add $800 in resale value, a well-made AI video can elevate a listing to the top of search results, make the car feel premium, and justify a firm asking price. Both are about presentation and perceived value — and both offer excellent ROI if used strategically.
In short:
WORK IN PROGRESS, will keep improving the guide, welcome your questions, ideas, suggestions.
I shared several of my early tests in that thread. While not perfect, they already look promising, so I decided to write a more detailed post on how to create these kinds of videos if anyone else is interested.
I truly believe that, if used correctly, these tools can help dealerships increase their sales. Videos are simply more engaging and leave a stronger impression on people browsing cars online. Because of that, I think they offer a clear advantage and are very likely to become popular soon.
Right now, the number of videos in car listings is still low mainly because the process is slow, tedious, and often produces poor results. If there was an easier way to make them, they would quickly spread everywhere. This post focuses on creating videos easily, with tools that are available to anyone and that are either free or inexpensive.
1. Start with Great Pictures
To make these videos, all you need are good-quality pictures. I cannot stress this enough: the better your source photos, the better your final video will look.
If you take poor photos, compress them, and then upload them again, the quality will degrade at every step. But if you start with clean, high-resolution images and keep them that way, your viewers will have a hard time telling whether it’s real or AI. The goal is for them not to notice any AI at all.
The great thing is that taking pictures is easy. Anyone can take a decent shot. That’s not the case with video. Walkarounds of cars often look shaky or unprofessional, and they can actually take away from the car instead of adding value. Either do a great walkaround or skip it completely.
By using AI, we can now create perfect walkarounds and cinematic movements that would normally require expensive equipment and a full setup. That alone is a huge benefit. With these AI models, we can take our high-quality pictures and turn them into professional-looking clips that truly showcase the vehicle.
2. The Tools You Need
I believe the best car sales videos include a host — usually a beautiful girl who presents the car and grabs attention from people scrolling fast. “Who is that beautiful girl next to the Ferrari?” That’s exactly the type of reaction we want.
The good news is that you don’t need to hire a model or a real person. It’s all AI.
The first thing we’ll do is generate our model. For this, you can simply use ChatGPT. Ask ChatGPT to generate the picture of the model however you want her to look. You decide her clothing, hairstyle, and overall appearance. Make it detailed. You can reuse the same model for multiple cars — she’ll become your own virtual influencer, always ready to present your next car.
Now that you have your female model and great photos of your car, it’s time to generate the first keyframe. Basically, we’re going to combine the model with one of your car pictures.
To do this, we’ll use an AI model called Nano Banana. It allows you to use multiple inputs (for example, your AI model and a car image) and create a single, combined output.
So here’s the idea: give Nano Banana the female model you generated with ChatGPT and one of your car pictures, and then add a prompt like:
It might take a few tries to get the right image, but once you do, you’re ready for the next step.
Generate Your First Clip
Next, we’ll use an AI model that turns an image into a video. There are several available, and you can find the best-performing ones on this list:
Image-to-Video Arena | LMArena
For my tests, I’m using Veo 3.1, as it’s currently giving the best results. However, this changes fast, so if you’re reading this later, check that leaderboard for newer models.
Once you’re set up, generate your first clip. You’ll usually want to make multiple clips of the same scene to add variety. You can also generate clips without the female model, for example, to show the car’s interior.
At this point, it’s important to understand prompting — how to describe what you want the AI to create. The clearer and more specific your description, the better the results.
A Few Words About Prompting
Prompting is basically giving instructions to the AI so it knows exactly what you want. A vague prompt will give you random results, while a detailed one will give you professional-looking output.
A bad prompt example:
A good prompt example:
A better prompt if you want to include your AI model:
Tips for writing good prompts:
- Always describe lighting (daylight, cloudy, indoor showroom).
- Mention camera movement (slow pan, orbit, steady shot).
- Include style or tone (cinematic, realistic, luxury ad).
- Add environment details (showroom, city street, open road).
- Use negative prompts to remove distractions: “no text, no logos, no watermarks.”
Once you learn to prompt properly, the quality jump is dramatic. The same AI model can go from average to amazing just by improving your prompt.
Adding Voice and Script
Next, you need a script — a short text that will be turned into voice. You can use ChatGPT for this part too. Give it some details about the car (make, model, year, key features) and ask it to write a friendly, conversational sales script.
Once you have the script, use ElevenLabs to generate the audio. It will convert your text into a realistic, natural-sounding voice.
Then, overlay the voice track on top of your clips. You can use CapCut to edit them online, or any other editor you prefer.
Fixing Lip Sync
At this point, you’ll have a nice-looking video, with different clips and your virtual model talking about the car. But there’s one problem: the lips don’t match the narration.
To fix that, use a lip-sync AI tool like Hailuo AI LipSync. Just search for it online. It’s simple to use — upload your video and audio, and it will synchronize the lips automatically.
Time and Cost
How long this takes depends on how many clips you create and how polished you want your final video to be. Here’s a realistic estimate:
- Creating base images and keyframes: a few minutes per image.
- Generating clips: 30–60 minutes total for 5 or 6 clips.
- Lip sync and voice generation: around 10–15 minutes.
- Editing and assembling: 1–1.5 hours.
So a complete video, with multiple clips and proper editing, usually takes 2.5 to 3 hours total.
As for cost, it depends on the models you use:
- Veo 3.1: around $10–20 per video depending on length and retries.
- ElevenLabs: a few dollars per project.
- Other models: most are free or inexpensive.
In total, expect about $20 in AI costs per video. If you add an additional (you also need time to generate the clips with ai) 1.5 hours of editing time, a finished video usually costs around $60–100, depending on who edits it and local rates. The time and costs to edit are the main costs., not the ai models.
That means for under $100 and around three hours of work, you can produce a professional-quality sales video that looks like it came from a full production team.
When an AI Video Makes Sense
Much like detailing, video content becomes more impactful as the value of the vehicle increases. Buyers spending $20,000 or more expect polished presentation. A professionally crafted video, even if AI-generated, signals effort and care — the same way a freshly detailed car does.
Here’s a quick guide:
| Car Price Range | AI Video ($100) | Detailing ($200 avg) | Traditional Pro Video ($500+) | Makes Sense? |
|---|---|---|---|---|
| Under $5,000 | ❌ Not worth it | ❌ Only DIY | ❌ No | Skip both. Basic photos and cleaning are enough. |
| $5,000–$10,000 | ⚠️ Maybe, for rare trims or fast sale | ✅ Good ROI | ❌ Too expensive | Focus on detail; AI video only if you want a slight edge. |
| $10,000–$20,000 | ✅ Worth it | ✅ High ROI | ❌ Overkill unless show car | AI video + detail = strong combo to stand out. |
| $20,000–$35,000 | ✅ Strongly Recommended | ✅ High ROI | ⚠️ Only for rare/sport cars | AI video pays off; pro video if budget allows. |
| $35,000+ or Specialty | ✅ Definitely | ✅ Mandatory | ✅ Worth the investment | Use everything: AI video, pro detail, pro photos if possible. |
Why AI Video Has the Edge
A traditional car sales video produced by a local videographer can cost anywhere from $500 to $1,500, depending on quality and editing. That’s perfectly justifiable for a $60,000 exotic — but probably overkill for a $15,000 sedan.
AI bridges that gap. For under $100, you can create a virtual host, insert them into polished clips, sync the voiceover, and produce a clean, modern presentation — something that normally requires a videographer, model, gear, and studio.
Tie-In to Detailing
Just like a $200 detail can add $800 in resale value, a well-made AI video can elevate a listing to the top of search results, make the car feel premium, and justify a firm asking price. Both are about presentation and perceived value — and both offer excellent ROI if used strategically.
In short:
- If the car is worth under $10k, focus on photos and detailing.
- From $10k–$20k, add AI video to stand out and sell faster.
- At $30k+, combine detailing, AI video, and professional photos for a full-stack presentation.
WORK IN PROGRESS, will keep improving the guide, welcome your questions, ideas, suggestions.
Last edited:
