In 2025, can AI become our social media assistant?

We’re coming to the end of a huge year for using artificial intelligence in marketing. There's been massive progress, new releases, summits, ‘world firsts’ and a lot of discussion. 

And, because of all that talk, you’d be forgiven for thinking everyone else has it all figured out when it comes to using AI in content creation. 

Figuring out what makes good content, and what works as part of the creative process, takes a while. 

Not quite. According to a benchmarking report from Influencer Marketing Hub, 69% of marketers have integrated AI into their operations. So far, so good. But only 34% have seen significant improvements, and 17.5% have experienced setbacks. 

As with anything moving at such speed, there’s a lot more bark than bite going on, and figuring out what makes good content, and what works as part of the creative process, takes a while. 

Above: AI industry and human creativity can combine to maximise workflow and output. 

This year we set out to crack how AI fits into one of our main workstreams; the monthly social content calendars we create for many of our clients. Our priority was always to ensure creativity stayed the responsibility of our creative teams, and to explore the supporting roles that AI could play. 

The first role is the one of workhorse. Basically, using AI to do tasks that require manual effort or repetition, leaving the team more time for creativity or strategy. This includes things like adapting copy written for one social platform to suit another (an efficiency for clients with stretched budgets). Working with one of the ChatGPT plugins that have sprung up, you can build this process right into your content calendars. 

You keep that ownership of the creative idea, execution or tone while AI takes out some of the legwork. 

New releases from the likes of Adobe Firefly and Runway can also help reduce certain creative tasks to a fraction of the time, using generative fill and retouching tools. And, brands (or their agencies) can start to train ChatGPT to read their brainstorm transcripts and slot key info into their briefing templates, taking out the admin side and leaving it ready to be finessed by, well, people. 

You keep that ownership of the creative idea, execution or tone while AI takes out some of the legwork. 

Above: Images from eight&four’s Coach AI tool which has been trained on clients’ products and brand worlds.

The second role is one more like a golden retriever; using tools with the ability to learn and execute with style and, most importantly, time and energy beyond the capacity of our creative team. 

The best first hand example of this I can offer is eight&four’s use of stable diffusion technology to build our proprietary tool, Coach. We have trained Coach on our clients’ products and brand worlds, meaning we can prompt it to create videos and images we’d never usually be able to devote the resources to accomplishing. 

Scale is important – larger volumes of content mean more opportunities for efficiency, whereas some smaller projects end up better done the old-fashioned way. 

A full shoot for a reactive post around the Northern Lights or the release of Wicked isn’t feasible, but Coach helps to rapidly produce something that feels true to brand, thanks to its training on hundreds of brand assets, and oversight from creatives with deep product knowledge. 

We've had some trial and error making this part of our day-to-day. We've found ways to save time, but found sometimes we're putting square pegs in round holes, such as integrating AI automations directly into Excel content calendars. It's why we're now building a bespoke workspace, platform12, to house content calendars. Scale is important too – larger volumes and variations of content mean more opportunities for efficiency, whereas some smaller projects end up better done the old-fashioned way. 

Above: AI can act as a golden retriever', with the ability to learn and execute with style.

What will this mean for the future of creativity on social in the coming year? It will become more common for those workhorse tasks – like asset iteration, copy adaptation, and translation – to use AI. This brings with it advantages in tailoring content across markets, social platforms and to users – including greater personalisation based on user interests or behaviour. There’s lots of potential here, but only where human nuance is still employed. If your agency is producing German content for the German market, there’s great danger in ChatGPT being your only German team member: audiences will notice. 

Brands need to focus on quality output and smart uses.

GenAI will continue to become more accessible, not just in creative programs, but built right into social platforms too. Meta is encouraging users to generate AI imagery within Facebook and Instagram with example prompts like ‘dolphins having a business meeting’. This isn’t something most social users are rushing to try out yet (why would they?), but if anything, it feels as if features like this, with rubbish outputs, are more likely to generate backlash or fatigue rather than enthusiasm or acceptance. 

Brands need to focus on quality output and smart uses to avoid this potential negative sentiment.