Prompt Engineering: Multimodal Prompting

We consume sound while scrolling visuals. We process words while absorbing colours and composition. We don’t live or think in a single format. As humans, we’re naturally 𝗺𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹. So why should our prompts be any different?

As I continue my learning through Google’s Prompting Essentials Specialisation, the power of multimodal prompting has become clear. It’s not just about asking in text, it’s about giving your AI tool more to work with.

What is 𝗺𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗽𝗿𝗼𝗺𝗽𝘁𝗶𝗻𝗴?
It’s when you combine different formats like text, images, audio, video, or code within a prompt to improve the quality and relevance of your output. It works best when you’re trying to translate or synthesise across formats.

For example:
You need to design a teaser for a joint event between two brands. Instead of just describing them, you upload their logos and brand colors, then ask the AI to generate visual directions for social media. The result? A much more aligned and on-brand output.

As a designer, this feels intuitive. We know visuals carry meaning text cannot. And great ideas often emerge from the 𝗶𝗻𝘁𝗲𝗿𝗽𝗹𝗮𝘆 𝗼𝗳 𝗳𝗼𝗿𝗺𝗮𝘁𝘀, not just the content itself.

The key is understanding what modalities your AI tool supports and using them with intention. Whether you're building brand assets, drafting storyboards, or generating product concepts, the more complete your input, the more refined your output.

Better prompts = better outcomes.