Instead of fully AI-generated ads that may lack a human touch, a hybrid approach is more effective. Humans provide the core creative asset (e.g., a short video), and an AI agent then automates the testing of different copy, edits, and targeting to optimize performance at scale.

A single shoot day can be maximized by meticulously time-blocking the production of different assets. For example, shoot a complex launch video in the morning, a simple talking-head explainer after lunch, and then use the remaining time to film several direct-to-camera ads for testing.

Don't just copy a competitor's ad. Instead, deconstruct multiple great ads into their core components. Then, reassemble a novel ad by combining a hook from one, a visual style from another, and a call-to-action from a third.

Instead of commissioning several unique videos from a creator, ask for one main video with multiple different hooks (opening lines/scenes). This gives you more ad versions to test for a fraction of the cost.

The reason a customer "needs" your product is subjective. Instead of a one-size-fits-all ad, create multiple versions that speak to different core buyer motivations. One ad might appeal to logic and data, another to time savings, and a third to team efficiency, ensuring you resonate with a broader audience.

While AI image models create high-fidelity ads, generating variations is costly. A cheaper, faster approach is building ad templates as code (e.g., React components). This allows for creating thousands of text and layout variations for free, enabling rapid testing of messaging before investing in polished visuals.

Instead of shooting one version of an ad, film the same core creative but with five or six different opening lines (hooks). This is nearly free to do on set and provides a wealth of testable assets to discover which pain point or message resonates most with your audience, dramatically improving campaign performance.

Start paid media testing with high-level message categories, or 'avenues' (e.g., 'designed by experts'). Once data shows which avenue resonates, drill down into minor variations, or 'cul-de-sacs' (e.g., 'handpicked by experts', 'backed by experts'). This structured approach prevents wasted spend on testing random copy.

Instead of writing a message on a sign and filming, record yourself holding a blank sign. This creates a reusable template. You can then use editing software to digitally add different text for each new piece of content, maximizing output from a single shoot.