How we’re using generative AI to work more efficiently

This summer, we were lucky enough to hit pause on the day-to-day, get together and dive into AI Fundamentals training with the incredible team at Spark AI.

The session was a fast track through what generative AI actually is, how it works, and why it’s only as good as the humans prompting it.

Spark’s approach stripped away the rocket emojis, and showed us how to make AI feel like a co-worker rather than a threat.

Prompting like a pro

The biggest takeaway from the session? Prompts are everything.

A simple “Can you make this sound better?” will not result in anything but a smattering of em dashes, a list of synonyms and a yawn-inducing samey-ness to your copy.

Spark compared generative AI to briefing “a super-intelligent intern on their first day.” Give AI some context, tone, and structure and suddenly your output stops sounding like the internet and starts becoming something usable (albeit, still with tweaks needed).

Crafting your own GPTs is another handy tool on offer. This functionality acts as your own personal prompt library, which can be used again and again to whittle away at some of those daily tasks.

Robot fact-checking

Just like “a super-intelligent intern on their first day”, if generative AI doesn’t know the answer to your question, it’s liable to try and blag it. That’s why it’s imperative that anything that comes from AI is scrutinised and fact-checked by a human safety net before printing or posting. Not just for inaccuracies, but also for repetition, pace and anything that doesn’t sound like you.

It’s also worth remembering that all AI models are trained on existing content. So, it’s impossible for ‘as-is’ generated copy or images to be truly original. Therefore they should never be published without significant human input.

AI as a collaborator

As part of the workshop we got hands-on with tools such as ChatGPT and Ideogram, with a series of creative challenges. In less than an hour, we had campaign ideas, copy and visuals. The point of the exercise wasn’t perfection, it was proving how fast and fluid creativity can be when humans lead and AI is there to support.

We were reminded that AI isn’t for churning out copy or images. It’s a starting off point, a brainstorming machine, a collaborator who is there, ready-and-waiting, when all your colleagues are busy.

Protecting your originality

It’s important never to feed an AI tool client names, financial information, internal docs, or unreleased campaigns. It’s best to assume anything you type could be stored or reviewed by the platform.

Investing in a version of generative AI that offers enhanced security (so your content won’t be shared), admin controls and a shared workspace so your team can work collaboratively with each other as well as AI is also a good plan.

Integrated AI

AI isn’t here to replace our creativity, but it can help to spark it. With the right prompts and a human at the helm, it can help us work faster, think broader, and stay ahead of the curve too.

After a few hours, our takeaway was simple: the best content in the future won’t be powered by AI. It’ll be powered by humans who know how to use it.