AI talking head ads

AI talking head ads are videos in which one AI-generated person speaks straight to camera, either seated at a table or desk or sitting in a car, and recommends or explains a product. The public examples from X below cover both settings; this page sets out how to choose between them, pace the script and make the clip with AI.

These are public examples of AI ads from creators on X. They were not made on or for Anglr. Click any ad to see the original post.

Two settings, one format

An AI talking head ad is one person speaking to the camera. The examples here come in two settings.

Seated Car
What is visible Person at a table or desk, home or studio Person in the driver's or passenger seat
Camera Steady, chest-up Phone on the dash or held up, slight shake
Product Often held or on the table (a drink, a snack) Usually mentioned; held up briefly if at all
Ratio in the examples Mostly vertical, some landscape Vertical, some letterboxed
Feels like A recommendation or explanation A quick, unplanned update

Pick seated when the product needs to be shown and handled. Pick the car when the script is a personal opinion or a rant and the product is easy to name.

Why brands use talking heads

A face speaking to the viewer is the most direct way to deliver a script, and it is fast to version: change the first line, keep the rest. It is also the format most scripts are first written for, so it is often where testing starts.

Beat list for a 20 to 30 second clip

  1. 0–2 s, the first line. Said straight to camera, no greeting. Captions on from the first word.
  2. 2–8 s, the problem. What the speaker or the viewer keeps running into.
  3. 8–15 s, the product. What it is and the one reason it fixes the problem. Seated: hold it up. Car: name it clearly.
  4. 15–22 s, the proof. A specific detail: how it is used, what changed, what makes it different.
  5. 22–28 s, the ask. One direct line telling the viewer what to do.

Pacing the script

Talking heads lose the viewer when nothing changes on screen. Plan a visual change every few seconds: a punch-in, a cut to a slightly different angle, a product insert, a text overlay.

Write one idea per sentence. If a sentence needs a comma to breathe, it can usually become two. For opening lines, see ad hooks; for a template with a shot column, see the UGC script template.

Framing and length

Make it vertical 9:16, eyes in the top third, captions below the chin. In the car version, keep the window behind the speaker bright but not blown out, and keep the seatbelt and headrest consistent between shots. Plan 20 to 30 seconds for the main cut, and test shorter versions built from the first line, the product beat and the ask.

Make it with AI

  1. Generate one character reference and one setting reference (the table or the car interior).
  2. Generate the voice for the full script first, so timing is fixed.
  3. Generate each shot from the references with lip-sync to its part of the audio. Vary shot size between cuts.
  4. Generate product inserts separately from a real product photo.
  5. Add captions and cut.
vertical phone video, [character description] sitting in the [driver's seat of a car / at a kitchen table], talking to camera, [emotion], natural daylight, chest-up framing, [product] in hand

Lip-sync and character notes are in the Higgsfield and HeyGen guides.

Compliance checks

  • Spokesperson claims. The speaker is AI-generated; don't present them as a real customer, employee or professional.
  • Expert framing. White coats, scrubs and job titles suggest expertise, so check the platform rules on expert claims; see AI expert ads.
  • Driving. In the car version, the person should not appear to be driving while filming.
  • AI label. Follow platform rules: AI ad labels on Meta and TikTok.

The base version of a winning script

The talking head is usually the version the winning script was written for, so it is the reference for every other format. Once it is made, the same lines can be split into a podcast, spread across a routine, or turned into text over a brand film. One winning script, made in every ad format covers that process.

Who this is for

  • Independent AI creators and editors. You make the ads yourself. You know your work performs, so you'd rather be paid on what it does than on a flat fee.
  • AI agencies and studios. You run a team of editors. You write the brief, your team makes the versions, and you're paid on performance because you know your team's ads scale.

Not for you if

  • You only want a flat fee.
  • You're new to editing and not yet confident your ads will perform.
  • You want every shot spelled out for you.
  • You make one style and won't try others.

FAQ

What is an AI talking head ad?

It is an ad where one AI-generated person speaks directly to the camera for most of the clip, usually recommending or explaining a product. The face and the spoken script do most of the work.

What is an AI car talking head ad?

It is the same format filmed as if the person is in the driver's or passenger seat with a phone on the dash or held up. The car makes it look like an unplanned recording between errands.

What is the difference between an AI talking head and an AI avatar ad?

An avatar tool usually gives you a stock presenter reading your text in a fixed setting. A made talking head ad starts from a character, setting, script and edit chosen for that ad, with cuts, captions and product shots added.

How fast should the script be?

Write short sentences and leave a cut at least every few seconds. A talking head with no change in framing for a long stretch is easy to scroll past.

Can an AI talking head present as a doctor or expert?

Expert claims need extra care. An AI-generated person should not be presented as a real, qualified professional. See the AI expert ads page for the checks.