Unlocking AI's Potential: 10 Tips for Elevating Your Production Process

Leveraging AI in production has recently emerged as a fundamental strategy for brands looking to stay ahead of the curve. However, diving into AI without a clear strategy can lead to pitfalls and missed opportunities. To help you navigate this terrain, we've compiled ten essential tips to help you harness the power of AI for your brand’s next campaign production.

‍

1. Brief your team on the risks ahead of time.

Before starting an AI-driven project, it's important to ensure your team is well-versed in the potential risks and regulations. AI isn't a one-size-fits-all solution, and aligning internal stakeholders on these details can prevent unwelcome surprises down the line.

‍

2. Don’t overpromise.

As tempting as it might be to promise your team the world, it's crucial to manage expectations realistically. AI's unpredictable nature means outcomes can vary, so it's best to underpromise and overdeliver rather than set unrealistic expectations.

‍

3. Craft a detailed creative brief (and be prepared to iterate).

The more you can flesh out your creative brief, the better equipped your team will be to prompt AI. This can even include sketches or guidelines. When working on one project which included an AI-generated alpaca, we included character rules such as: extremely fluffy, never shorn, doesn’t make human poses, wears black sunglasses, obeys laws of physics, is expressive but doesn’t overact.

‍

‍

‍

Make sure the team is open to iteration as you explore AI's capabilities—adjusting the brief based on what's possible.

‍

4. Foster collaboration between tech and creative teams.

Close collaboration between your AI tech team and art directors is key. Afterall, the person proficient in the technology platforms is not always the person with the creative vision. Having these teams in lock-step will ensure the project's effectiveness and productivity.

‍

5. Set the stage.

Provide AI with abundant details, stimuli, and context upfront to give it the best chance of success. More information means better results, so don't skimp on providing the necessary inputs.

‍

6. Leverage ALL available technologies.

Don't limit yourself to one AI technology. Experiment with different platforms to find the one that best suits your project's needs and desired style. 

‍

When working with Varonis, we tried Midjourney, DALL-E 3, Stable Diffusion XL, Adobe Firefly, and Microsoft Designer to get a feel for the control and variability of each tool. The more we generated, the more we realized that Midjourney and MS Designer worked best for Varonis’ content. However, for our Joey project, DALL-E 3 was the best match.

‍

7. Start wide, then refine.

Begin with a broad approach then narrow in on what works best, fine-tuning your strategy iteratively. Our philosophy with AI is “Generate a lot! Throw away a lot!” We start with simple prompts, documenting them as we go. We share prompts and outputs often to dial in on the aspects that are working. Then we double-down on those and generate a whole lot more.

‍

8. Remember, prompting is an art form.

Crafting compelling prompts requires attention to detail. The way in which you ask AI for something—and then follow up with subsequent asks—is a very specific skill that can be honed over time. 

‍

9. Take it one detail at a time.

You might not get everything you want from your first prompt. And the fewer things you ask the technology to focus on, the less overloaded it will be and the more accurate it can get with each element. Start with the essentials and incrementally layer in additional aspects to get the result you’re looking for.

‍

When we worked with Varonis, we needed to generate an image in which a cake was just right. We started there and then went on to the next element—a pair of oven gloves. Then we asked AI to put those gloves in the picture with the cake. And voila (incrementally).

10. Lean into post-production.

AI might only get you 80-90% of the way there. So post production is where the magic unfolds. Backgrounds, colors, logos, etc. can all be refined, layered, or added in this phase. This means you can spend less time agonizing over perfection during the generation phase.

‍

Bonus Tip: Explore motion for added impact.

Motion adds an extra layer of engagement and can breathe life into your creations. AI tools are great for adding movement to static images and Image-to-Video or Text-to-Video AI tools can bring concepts to life with subtle or guided movement in up to 16-second generated clips.

‍

Leveraging AI in production presents huge opportunities for brands. By following these tips, your team can navigate the complexities of AI-driven projects with confidence, efficiency, and success—while having a little fun along the way.

‍

Want to chat more about how to leverage AI for your next project? Drop us a line. 

‍