Guides · 11 August 2026 · 2 min read

Adding a call to action and text to an AI video

Why the sign in your generated shot reads as nonsense, and the reliable way to put your offer and your address on screen.

By The Crewnema team

An advert that does not ask for anything is a nice film. The last thing a viewer sees should say what to do next, and that is where AI video most often goes wrong, because image and video models are poor at lettering.

The problem with text in a generated shot

When a model draws a shop sign, a price tag or a laptop screen, it draws something shaped like words. Sometimes it is close. Often it is a few real letters and some invented ones, and viewers notice. Do not rely on any model to spell your name, your price or your website inside the picture.

Do it in three layers

1. Say it. Put the offer in a short spoken line in the script, such as “Order by eight and we deliver today”. Spoken words are generated as audio, not drawn.
2. Print it on the end card. The closing card carries your call to action as real text set by Crewnema, not drawn by a model: an address, a phone number, a website or a code.
3. Keep lettering out of the shots. In shot descriptions, avoid signs and screens with readable text, or describe them as blurred or out of focus.

Write a call to action people can act on

• Start with a verb: order, book, visit, call, join.
• Give one next step, not three.
• Make the destination short enough to remember or type: a website or a number, not a long link.
• Say why now when it is true: a date, a limit, a festival.

Instead of: “We hope to see you soon at our shop.”

Try: “Book your first class free this week. Visit fitwellstudio.in.”

The second says what to do, where, and why now. The first asks for nothing.

For the whole workflow, see how to get high-quality AI video.

Make your first video

Crewnema is in private beta and lets businesses in a few at a time. Request access and we will write when yours is open.