It’s a little choppy for marketers out there at the moment, isn’t it?

GenAI has already been preemptively charged with the murder of our industry’s *consults notes* copywriters, graphic designers, video editors, media buyers, social strategists, market researchers, community managers, photographers and video directors. And, awkwardly, every department of my own social and content agency.

Even Sam Altman, CEO of OpenAI, has been giving us a kicking. Just a few weeks ago he predicted that 95% of us are going to lose our jobs to the GenAI machine within five years. Last week McKinsey revealed they’re paying marketers to leave. Will we all be victims to the onward march of AI?

Honestly, none of us know. Not even Altman. GenAI (AI’s shiniest offspring) has spawned some of the most engaging technical innovations in a decade. A chatbot that can look at a picture of my broken bike and tell me how to fix it? That can make me rap like Kanye? Take my photo and make me a Disney princess?

But trying to tease apart the extraordinary novelty, from the all-important utility, is harder. After 12 months spent harnessing GenAI for my business, I’ll share my thoughts.

First, it’s important to understand clearly how GenAI, and all its spawn (ChatGPT, DALL-E, Midjourney, Stable Diffusion, Claude, Runway, Google Gemini etc) actually work.

And that’s simply, prediction. In the case of copywriting, ChatGPT works by predicting the next word that’s most likely to follow in a sequence, considering all the words that came before it. In the training stage it digests a massive amount of data and observes how words and phrases typically follow each other. With each advancement it gets smarter, understanding context, and creates more accurate, coherent and compelling copy.

These complex statistical methods yield extraordinary results, but the probabilistic nature of the tools causes issues for marketers.

1. Specificity. GenAI is trained on broad, global patterns and struggles with tight windows of specificity, such as communicating a client’s product or service in a highly accurate way, with a specific tone of voice, in a tightly controlled brand world, and carefully considered objectives.
2. Creativity. GenAI is trained on data that already exists, so it’s returning the average of what the world has created to date. Not exactly a winning formula for original thinking.
3. Legality & ethics. Marketers can get undone by a legal team’s concerns around copyright, consent, biases and ethical use cases, putting the reins on GenAI innovation.
4. Overreliance. AI is a great tool for the lazy, and we’re biologically designed to chase efficiency, so it’s critical marketers don’t become over-reliant on - and indistinguishable from - GenAI. Then we’ll definitely lose our jobs.

So what is it good for? In short; productivity. That’s it. We’re seeing 10-30% time savings across marketing tasks - and up to 80% on some specific tasks. This productivity is the central benefit to marketers, and one which must be harnessed to stay relevant in the industry.

As for productivity boost, these are the key areas we’ve seen AI help in our business.

Copywriting

ChatGPT, Claude, Gemini and their copywriting cousins are causing understandable nervousness.

These tools are great for: speeding up market research, sourcing content in 10% of the time of desk research - Perplexity AI is especially useful for getting explicitly-sourced material; functional asset copy - as part of platform12, our proprietary platform, we train our tool on the client’s copywriting TOV and best practice assets, and let it deliver the volume at a much more cost-effective rate; and greasing the wheels of ideation/long-form content writing. No more blank paper: Flot.ai and Springboards.ai are two great tools to look at in this space.

What’s it not good for? Storytelling. These tools can do a great synopsis, they can’t (yet) tell a great human story, because they have no lived experience. Humans want to hear stories from other humans. We want emotional resonance to feel human connection.

Design & Video

Midjourney, DALL-E, Firefly, Runway - lots of different options here. In order to create a secure, closed environment for clients, we’ve developed a creative tool, Coach, using Stable Diffusion technology.

Where’s it winning?

1. Replacing product photoshoots. Last year a Diageo Marketing Director said the product photoshoot was dead. He was right. With Coach we’ve been undertaking international automotive photoshoots across five different models, 32 specifications and 100+ locations without leaving our desks, delivering cost savings of 80% on a traditional photoshoot model.
2. Accelerating illustration projects. Coach is trained on our mobile gaming client’s data so we can generate on-brand artwork for their brand worlds, saving a single person 4-5 hours a week on artworking stock imagery.
3. Enabling enhanced personalisation. Please don’t be Adobe and celebrate the “avalanche of content” GenAI will allow you to produce. But, some cost-effective localisation may suit targeted media buying - and help put out some ‘content deficit crisis’ fires.
4. Removing the budget guardrails from creative thinking. Take a look at OpenAI’s research paper on Sora – and view the surfers in the library. We are about to see an incredible leap in the minimum viable benchmark for creative marketing.

What does this actually mean for marketers?

Regardless of what happens next, the benchmark bar has been raised. If ChatGPT can write OK marketing copy, no agency can still afford to.

Secondly, it will cause widespread disruption to how agencies/marketing teams organise their marketing workbench. People will remain absolutely critical, but companies will look to balance human capital with investments in proprietary/partnered tech.

It’s critical that marketers get their hands on these tools and experiment with the precise areas where they can drive productivity for their business. But do it fast, because GenAI will wait for no man - or marketer.