Loading…

AI for ad images and video creative

6 minutes, 8 steps, workout 3 of 6 in this module

Image and video tools can now turn out dozens of ad visuals before lunch, and someone on the team will suggest replacing the photo shoot with them. Sometimes that's the right call, and sometimes it quietly costs trust, for example when the product in the picture isn't the one the customer receives. Knowing which parts of a creative to generate and which to keep real, and how to feed new versions into a proper test, is quickly becoming a normal part of running paid social.

What you'll be able to do

By the end of this workout, you'll be able to decide which parts of ad creative to hand to AI tools and which to keep real, and use AI to feed a creative testing loop instead of replacing it.

A first look: Where AI helps with creative, and where it hurts

AI tools are good at the parts of creative work that are repetitive: generating backgrounds and settings around a real product shot, resizing and extending an image to every placement, producing variations of a hook or headline, drafting a storyboard, writing a scratch voiceover for a test. Meta and Google both have generative features built into their ad tools that do some of this automatically. AI tools are bad at the parts that have to be exact: the product itself, text inside an image, hands, and anything presented as a real person or real customer. Keep those real.

Example: Green Garden Skincare shoots one clean jar photo. From that, the team produces a bathroom-shelf version, a gym-bag version and a morning-desk version with an AI image tool, plus 1:1, 4:5 and 9:16 crops. One shoot, nine test creatives.

What this workout covers

  1. 1Where AI helps with creative, and where it hurts
  2. 2Case: Which image gets the budget?
  3. 3AI moves the bottleneck from production to judgement
  4. 4Case: Corner Coffee's 15-second TikTok

A question from this workout: Corner Coffee's 15-second TikTok

Corner Coffee wants a 15-second TikTok for the new station location. The agency offers two routes. Route 1: an AI-generated barista avatar in a generated café, done in a day, no filming. Route 2: real staff film the counter on a phone, and AI tools handle captions, cuts, three hook variants and the vertical crop.

Which route do you pick, and why?

  1. Route 1, it's faster and looks more polished
  2. Route 2, because a real counter and real staff are the proof; AI speeds up the editing, not the truth
  3. Route 1, but add a small “AI-generated” label so it's honest
  4. Neither, TikTok needs a professional production company

Pick your answer first, then open the reasoning below.

Show the answer and the reasoning

Answer: B. Route 2, because a real counter and real staff are the proof; AI speeds up the editing, not the truth

A café sells a place and people. A generated avatar in a generated café removes exactly what the ad is supposed to prove. The editing work, captions, cuts, hook variants, is where AI tools save a day. The most common mistake is option c: the label makes it honest, but it doesn't make anyone want to visit a café that doesn't exist.

The other steps work the same way: you decide first, then see why each option is right or wrong.

How you practice here

  • 2 short concept cards
  • 2 multiple-choice questions
  • 2 campaign cases with real-looking numbers
  • 1 matching exercise
  • 1 guided calculation