Author: Ross Simmonds
Businesses that rely on social media marketing know that there’s rarely enough hours in the week to research your audience, plan content, publish and promote it, and subsequently engage with your followers about it.
Generative artificial intelligence has the potential to change that. But for all AI’s capabilities, it’s still too risky for brands to let it run your social media presence without active human guidance.
Over the last three years, I’ve worked closely with some of the top brands in the world to understand how AI can support their workflows and processes. I’ve spoken at conferences around the world on the power of AI and how marketers can use it to drive better business outcomes—outcomes that include:
More efficient and productive teams
More agile creative processes
More velocity for content production
Better return on investment on social media
And in this essay, my goal is to share all of that with you.
Let’s get into it:
Should you use AI for social media? Strengths and weaknesses
Social media and artificial intelligence have a long relationship. After all, AI and machine learning power the algorithms that determine what content shows up on your customers’ feeds. As digital marketers, it only seems logical to consider AI within our own workflows and systems for creating social media content.
But, how do you do that? What are AI’s strengths and weaknesses when it comes to social media? And, more importantly, how do you match its strengths to your brand’s marketing needs?
The first thing I want to specify is:
AI’s efficacy depends on the particular use case. Your success with AI for social media depends on how well you can blend its strengths with your own, as the human expert, marketer, or business owner.
I’ll also add that in a few years, it’s very possible that everything I’m saying you shouldn’t use AI for today will be 100% feasible to use AI for at that point (but, I’ll probably have updated this content by then).
However as it stands right now, I avoid using AI for the following social media tasks and advise that others avoid it for these use cases a well:
Strategy development
Social account management
Unedited long-form social content
Unsupervised customer service
Why?
ChatGPT and most LLMs currently do not have the capability to research audiences, conduct keyword analysis, analyze communities, or dive into the qualitative and quantitative data associated with your email list and use that to figure out a plan. These are still tasks that require human intuition, experience, and common sense (remember, ChatGPT is just putting words together based on relationships with other words—it doesn’t actually do any ‘thinking’).
It’s also quickly becoming obvious to customers when a brand creates and distributes AI-generated content. The best brands include a human layer in the process before publishing to ensure that the content is actually what human customers might want to see. Not to get redundant, but remember, AI does not shop for products or services so it’s really unlikely to generate the best (unedited) content for those purposes.
But that doesn’t mean you should avoid AI at all costs when running your social media presence.
AI can be effective for… | Because… |
| AI handles a variety of lightweight tasks (e.g., short Q&A responses, brainstorming ideas) quite well. It can also perform well as part of a workflow (rather than to simply generate content that users will see) because of the human oversight inherent in bigger tasks/projects. |
So, how can you do these things? Let’s dive in.
Prompt engineering: Tap into the full potential of AI for social media
What’s the best way to create content for social media? It’s the combination of great prompt engineering and creativity.
You’re probably familiar with striving for creative excellence, but prompt engineering might be a newer concept for you. Here’s how it works:
Prompt engineering is the process of crafting and refining the instruction or query you feed to a generative AI tool to improve its output. Prompt engineering probably won’t get you ‘perfect’ results, but it can vastly improve your starting point.
When you’re creating content for social media, it’s not enough to write a simple (or ambiguous) prompt like:
Create a LinkedIn update promoting my brand.
You need to optimize your prompts for the content you’re looking to generate.
Here are the four most important things to remember when crafting AI prompts for social media assets to use in your marketing material:
Define your goal or objective: Your prompt must clearly define what the AI should produce (e.g., the desired format, target audience, tone of the content).
Set the context: Provide relevant background information to help the AI model grasp your expectations, leading to more accurate and pertinent responses.
Offer examples and guidance: Include examples in your prompts to give the AI a clear template to follow, especially when you want it to adhere to specific structures or guidelines.
Continuously iterate and refine your prompts: Just like any solid communication strategy, prompt engineering is iterative. Test different variations, evaluate the results, and refine your prompts to align them with your evolving objectives.
Here’s an example of how to create a great prompt using the techniques I mentioned above:
The prompt above excels at providing the AI with:
The role it needs to play (i.e., copywriter)
Which creators to take inspiration from (i.e., Ann Handley and Seth Godin)
A clear framework to use (i.e., AIDA)
A series of examples to use for inspiration
And then at the end of the prompt, it tells the AI exactly what it wants:
The results?
Social media posts that attract attention, align with the ideal voice, and (with some editing) can be scheduled to go live immediately to drive some engagement and buzz.
AI for customer engagement on social media: Proceed at your own risk
Replying to customers, fans, and followers is one of the most common AI use cases on social media.
When I first started in digital marketing, I worked in ‘social media newsrooms,’ where social media specialists monitored Twitter (now X) for posts about companies and replied to these posts in seconds.
Today, AI is tackling this task on behalf of some of the largest companies in the world. Rather than outsourcing this type of work to a call center or social media communications firm, many organizations use AI-generated agents to manage replies, responses, and criticism.
Does it always work? Not a chance. There’s a lot of examples and press about AI-generated responses that miss the point of the customer's (sometimes even influencers) original social media posts.
This is why it’s important to be careful with how far you go in deploying this technology.
Can it help you write replies faster? 100%.
But, should it be 100% automated? That’s a risk and you have to decide whether you’re willing to take it or not.
AI for social content production and distribution
As technology advances, so does its impact on industries and professions. Do not overlook the ways in which generative AI influences how designers, writers, videographers, editors, etc. do their jobs.
In recent years, there’s a growing trend of using AI tools to assist in content production tasks, such as writing, designing, animation, editing, and even recording video (as I’ll show you later).
For example, Midjourney (shown above) is an independent research lab that provides AI-generated images based on text prompts, similar to other AI image generators, like OpenAI’s DALL-E.
As you can see from the image below, the text doesn’t always come out perfect (which is incredibly common with AI-generated images at this point in the technology’s maturity) but the output looks pretty good:
One of the use cases I’ve experimented with quite a bit on social media is using generative AI to create video content. Here’s an example of a video of me (kind of) and me:
I used HeyGen, one of my favorite AI tools for this. Can you tell which is human and which is AI-generated?
I’ve also used AI tools for content distribution, turning blog posts into tweets and crafting social media posts to promote podcasts. Tools like Distribution.ai (which is my own) make it possible for me to upload audio files from my podcast, Create Like The Greats, and then create LinkedIn posts and X content from that audio file in a matter of seconds.
This is what works for me and my business at this point, and while I recommend these tools, I encourage you to test and take advantage of free trials and demos to figure out what’s feasible (and just as importantly, sustainable) for your social media marketing goals.
Wix AI tools for social media
Starting with what you already have access to (or can access for free) is a great, low-risk way to experiment with AI.
Wix website owners can access a range of AI-powered features to expand their online presence. The AI image creator is a great example:
You can even use the Photo Studio feature to upscale resolution, erase an object, extend the image, remove the background, etc.
Additionally, one of the most common AI uses for digital marketers is generating copy for social media posts. AI tools, like ChatGPT, now give social media managers the ability to create this copy 10x faster.
As an example of this in action, instead of me thinking for an extra 10 minutes about this blog post’s title tags and meta descriptions, AI can do the heavy lifting. Wix’s AI meta tag creator makes it easy for me to generate this content in seconds.
While this feature is typically used for web content, you can set your og:description (social description) and og:title (social title) to mirror your title tag and meta description so that the AI takes care of even more for you. Of course, you should still review the outputs and seize any opportunities to further optimize it (with a CTA, for example) before publishing.
Add AI to your social media workflows: Considerations and suggestions
The opportunities to incorporate AI into your workflows as a social media marketer are significant. It can be a go-to for so many parts of your job, including your research, creation, and optimization processes.
But, here’s one area that most people overlook: AI as a brainstorming tool.
One of the best ways to avoid writer’s block is to treat AI like a creative partner. Start a dialog directly with your favorite LLM tool to discuss some of the things that you’re thinking about as it relates to possible content ideas. One way of doing this is to find a social media idea that works in one industry and then ask your favorite AI tool to use that for inspiration in creating new ideas.
Let’s use the HVAC company again as an example. I could take a post from Semrush (the SEO rank tracking tool) which talks about a marketers ‘dream 24 hours’ and ask ChatGPT to use it for inspiration on a ‘dream 24 hours’ post for an HVAC professional:
And in a matter of seconds, it responded:
“Why do we love working in HVAC? Sure, there are the technical skills, the challenging repairs, and the customer satisfaction... but let’s be real, we’re here for the good stuff: comfort and efficiency!
The dream 24 hours of any HVAC pro:
09:00 - Kicking off the day with a successful repair, leaving a customer’s home warm and cozy.
10:00 - Learning new energy-saving tips to share with clients.
11:00 - Installing a high-efficiency system that cuts energy bills in half.
12:00 - Getting a 5-star review from a happy homeowner.
14:00 - Diagnosing issues in record time with the latest HVAC tools.
16:00 - Wrapping up a commercial project ahead of schedule.
18:00 - Celebrating the day’s wins over coffee with the team.
🔥 BONUS: All service calls booked today get a 20% discount!
👉 Watch out: This offer is only valid for the next 24 hours!
Not bad eh?
While you’re probably excited about the potential of AI for your social media strategy, it’s very important to take a step back and assess your current social media plans.
Before you start signing up for tools, you need to understand:
How much time you can actively invest on this channel
Your potential budget for AI tools
The tasks that stand out to you as the most worthwhile/relevant for utilizing generative AI
That having been said, one recommendation I like to make (that almost every creator and business can benefit from) is to use an AI tool that schedules content to publish at optimal times. A lot of the social networks offer scheduling directly on their platforms, but tools like Hootsuite provide a depth of features that go above and beyond the scheduling functionality built into LinkedIn, Facebook, X, or any other network.
AI isn’t about perfect outputs—it’s about incremental efficiency gains for your social media workflows
“AI will do 95% of what marketers use agencies, strategists, and creative professionals for today.” — Sam Altman, CEO of OpenAI
That’s quite a bold statement, especially from my position as a marketer.
So, here’s my take: I’m not sure how close we are to a superintelligence that can do all the things we do. But, I do know that some of the tasks we focus on today will eventually be completely irrelevant thanks to generative AI. That’s not a prediction—it’s a fact.
My suggestion to social media marketers and managers is simply to experiment with the technology and look for advantages and opportunities. Look for ways to do more with less and see how AI can facilitate your workflows to be more efficient, but just as creative and human.
Ross Simmonds is the founder of Foundation Marketing, a B2B SaaS Marketing agency that works with some of the worlds most successful brands. He's also the author of Create Once. Distribute Forever: How Great Creators Spread Their Ideas and How You Can Too. Twitter | Linkedin