How to Make Product Ads With AI for TikTok and YouTube—For (Almost) Free

TikTok Shop, the e-commerce platform that lets users buy and sell things directly on TikTok, has moved $64.3 billion in merchandise in 2025, almost doubling from the previous year. The US alone was responsible for $15.1 billion of that, and most of that was driven by short, cheap, face-to-camera videos of someone holding a product and telling you why you need it.

Those videos used to require a person, a phone, decent lighting, and several takes.

Today, however, all you need is a product photo and three AI tools, most of which are available for free. Here’s a complete workflow, step by step, with no technical background, so you can start your marketing empire.

Step 1: Have a clean product image

Before you do anything else, you need to know what you want to sell. For this, you have a few options: choose something you’re passionate about, or just go ahead and find the most popular items on the TikTok store and download what you think will click.

Download a photo of the product you want to sell: a piece of clothing, an accessory, a gadget, whatever. If you sell a specific product or are affiliated with a company, use your own supplier photos. Then crop the image so that only the product is visible, with no model, no background noise, and no watermarks.

In our examples we chose this green top for TikTok (vertical format) and this Ledger crypto wallet for YouTube (horizontal format).

That crop is more important than it seems. The AI ​​will treat this image as the source of truth for the product, so the cleaner the reference, the more faithful the result.

Step 2: Place a model in your product

This step is important if you are promoting clothing and accessories as they require a more human touch.

Open ChatGPT and upload the cropped image. For this step you will need GPT Image 2 Declutter‘s own head-to-head testing, it beat Google’s Nano Banana 2 in photorealism and product fidelity, which is exactly what an AI-generated ad needs to not look fake.

Then imagine the setting you want for the ad, and turn it into a quick prompt.

You can use something like this: “Generate a vertical 9:16 photo of a Latina woman in her late twenties wearing this exact garment, posing for a casual smartphone photo in a bright apartment. Maintain every feature of the product exactly as shown in the reference image: shape, proportions, color, fabric texture, stitching, and fit. Do not re-design, color, or modify the product in any way.”

We have something like this:

Mix up the demographics (ethnicity, age, body type) to match who your target audience actually is. Change the setting in the same way: a gym for activewear, a cafe for accessories, a street corner for streetwear. For non-apparel products, replace ‘wear’ with ‘hold’ or ‘use’.

Do you want the model somewhere specific? Upload a second image of the location and ask ChatGPT to place the subject of image one in image two. Here our subject is on a space station, just because.

This trick also works with Google’s Nano Banana 2, which handles compositions well. Reve is a much cheaper alternative, but it often omits quick details and the accuracy of the product can suffer.

Step 3: Ask ChatGPT for the script – in JSON

Now you need a script for a 10 second video. Don’t write it yourself; let ChatGPT do the marketing thinking. Something like this can work, but be as detailed as possible in what you ask:

“Trap like a senior direct-response marketer. Write a 10-second script in English for a UGC-style video in which the woman in the attached image talks to the camera and sells the attached product. The copy should sound natural and spoken, engage the viewer in the first two seconds, mention that the price is $20, and close with the call to action ‘tap the shopping cart below.’ Run the script as JSON, formatted for Google Flow, with a timeline describing what’s happening on the screen, the camera behavior, and the exact dialogue for each segment of the 10 seconds.”

The JSON format is not a decoration. Video models, especially from Google, follow structured timelines much more closely than individual paragraphs, so you get the dialogue, gestures, and beats you asked for. One caveat: watch the output, as it can be so literal that if the timeline ends at second eight, the model may repeat an action to fill the remaining two seconds.

If you want, you can personalize the copy per platform. In the prompt, ask the AI ​​to say things like “The best top I’ve seen on my TikTok feed.” To switch social media, ‘tiktok feed’ becomes ‘on X’, ‘in my roles’ or ‘on Shorts’ depending on where it appears. The call to action also changes: the shopping cart works on TikTok, ‘link in bio’ fits Instagram, and ‘check the pinned comment’ fits YouTube.

Step 4: Generate the video in Google Flow

Go to Google Flow and select Gemini Omni, the model that Google launched at I/O 2026 in May. It generates clips of up to 10 seconds with native audio (meaning your model actually speaks the dialogue) and it accepts reference footage, which is the whole point here.

Google Veo works, and could arguably be better, but Omni is cheaper… and we like cheap.

Upload both files for reference: the generated image of your model and the cropped product close-up. Paste the JSON into the prompt box. Choose vertical 9:16 for TikTok, Reels and Shorts, or horizontal 16:9 for standard YouTube videos and pre-roll ads.

Now the money part. Flow gives non-subscribers 50 free credits per day, but Google’s supporting documents limit this to the Veo 3.1 models.

Omni in Flow (the model we recommend) requires a paid Google AI subscription. Plus ($7.99 per month) includes 200 monthly credits, Pro ($19.99) includes 1,000, and the two Ultra levels include 10,000 and 25,000.

However, there is a truly free backdoor: Google has made Omni available at no costs in YouTube Shorts and the YouTube Create app for users 18 years and older. And for reference, the developer API prices Omni at around $0.10 per second of video, about a dollar per clip.

Each Omni video features Google’s invisible SynthID watermark, identifying it as AI-generated. It won’t appear on screen, but platforms can detect it, so don’t plan on pretending the footage is real.

Here is the Ledger ad video we got. Keep in mind that video generation contains a lot of random components, as creativity is the key to making these models work. If you don’t like the first generation, try it a few more times.

If you notice irregularities that can be resolved later, you do not need to spend additional credits. There are free tools that allow you to cut parts of the video, change the lighting, color, etc.

For example, in this Tiktok ad, the woman repeats the call to action. We need to change that, and in the next step you’ll find out how

Step 5: Polish it with CapCut

Export the clip and open it in CapCut. Here you trim any anomalies (AI video still produces the occasional extra phrase or looping gesture) and cut the clip to exactly what you want before exporting it directly to your social accounts.

Subtitles are the only feature worth paying attention to. The animated, word-for-word caption styles that dominate TikTok are behind CapCut Pro, which costs about $7.99 per month or $59.99 per year, and the free automatic captions. Manual text remains free, so if you’re patient, you can type your own text.

The Tiktok ad ultimately looked like this after we adjusted the call to action:

Going down the rabbit hole

This workflow produces acceptable results, not agency quality results. Once it clicks, you’ll want more control: ElevenLabs for a consistent brand voice across dozens of videos, Kling for persistent AI avatars and motion control tools, Node-based workflows like ComfyUI for granular scene control, n8n for automation etc.

Each adds cost and complexity, but the basic pipeline above is enough to test whether a product sells before spending anything serious on it.

Selling through TikTok Shop doesn’t require much on your part: you just need to be 18 years or older, with government ID and bank details that match your registration, and approval will usually come in three days. TikTok charges a 6% referral fee on most orders. Promoting others’ products as an affiliate requires 1,000 followers to sign up, and full access requires 5,000 followers plus 30 days in the program.

YouTube has lowered the barrier even lower: in March, the Shopping Affiliate program opened with only 1.5 million euros for creators of the Partner Program 500 subscribers in twelve countries, including the US and Brazil.

Both TikTok and YouTube allow AI-generated promotional videos, but creators and advertisers must disclose how the content was created and the commercial relationship behind it. Below TikTok’s AI-generated content policymust be labeled realistic AI-generated images, audio or video; Advertisers running non-Spark ads must also enable the ‘This ad contains AI-generated content’ option TikTok ad managementwhile anyone promoting a brand, product or service must enable TikTok’s disclosure of commercial content setting.

YouTube similarly requires creators to select “Yes” under “AI use” when a video contains realistic content generated or meaningfully modified with AI, after which YouTube applies an AI label; sponsored, endorsed or otherwise commercially influenced videos must separately use the platform’s paid promotion revelation.

X’s Authenticity Policy prohibits synthetic or manipulated media when it is presented in a deceptive manner and is likely to cause widespread confusion, threaten public safety or cause serious harm, while advertising rules require that advertisements be fair, lawful and consistent with the product being promoted. It does not explicitly prohibit the use of AI-generated images, video and audio in promotional content.

But before you quit your job, let’s put your feet on the ground with a sobering number: Per Camille Moore, president of the marketing agency Insights into the third eyeOf the 803,500 TikTok Shop stores operating in the US last year, more than half recorded zero revenue. The tools are virtually free. Not the competition.