Artificial intelligence has been part of digital marketing for considerably longer than most people realise. Marketers have been using platforms powered by machine learning and automation for years, particularly when it comes to advertising, targeting, recommendations and analytics. What changed with the arrival of generative AI tools such as ChatGPT was that suddenly the technology became much more visible and accessible to everybody.
Now we can use AI to brainstorm content, write headlines, generate images, help with advertising, answer customer questions and potentially automate even more of our marketing. The opportunities are enormous, particularly for small businesses that don't necessarily have large marketing teams, but that also raises an important question: just because AI can do something for us, should we automatically let it?
For episode 262of The Monday Morning Marketing Podcast, we were delighted to be joined by Mari Smith, one of the world's best-known Facebook marketing experts, to talk about using AI in social media marketing properly, responsibly and, perhaps most importantly, without removing the human element that made social media valuable in the first place.
AI should be your assistant, not your replacement
One of the most useful ways small businesses can approach generative AI is to think of it as an extension of themselves rather than something that takes over the entire marketing process. Mari described it almost as having a second brain, something you can use to brainstorm, iterate and explore ideas when you're stuck.
Content creation is an obvious example. You may know that you want to create a series of short videos or Reels, but when you sit down to actually produce them, suddenly you can't think of a single useful thing to say. Instead of asking an AI tool to simply "write 20 Instagram Reels", you can give it information about your audience, products, services and objectives and ask it to help you identify the questions and pain points your customers are experiencing.
Better still, ask the AI to question you first. Tell it that you want to create a series of videos and ask what information it needs from you before making suggestions. This turns the process into more of a conversation and gives the tool more context to work with.
The important distinction is that you're using AI for ideation, not automatically copying and pasting everything it produces. Your experience, opinions and understanding of your customers still need to be part of the finished content.
If AI created it, should we say so?
As AI-generated images and video become more realistic, transparency is going to become increasingly important. Mari used the term "synthetic media" to describe content generated by AI and raised concerns about how easily people can now encounter images or videos without immediately knowing whether what they're seeing is real.
That becomes particularly uncomfortable when we're dealing with apparently real people, situations or events. The technology is advancing quickly enough that simply looking at something may no longer be enough to determine whether it actually happened.
Platforms such as Meta have introduced labels and watermarks for some AI-generated content, and Mari's view is that this type of transparency should become much more standard. If AI has played a significant role in creating something, telling people doesn't diminish the content. In many cases, it actually helps protect the trust you've built with your audience.
Interestingly, the rapid growth of synthetic content may make something we've talked about in marketing for years even more valuable: authenticity. Real people, real opinions, genuine experiences and human interaction could become increasingly important precisely because audiences are going to encounter so much content that isn't real.
AI can help with advertising, but don't hand over the keys just yet
AI is already deeply embedded within Meta's advertising tools, whether advertisers consciously think about it that way or not. It can help generate alternative headlines and primary text, enhance creative assets, adjust placements and use Meta's Advantage+ tools to help identify audiences.
Some of these applications make a lot of sense. If you're advertising a physical product, for example, being able to take your own product photograph and experiment with different backgrounds or presentation styles can make the creative process considerably easier. Likewise, allowing the advertising platform to optimise placements based on performance and available inventory can reduce some of the manual work involved in campaigns.
Where things become more complicated is audience targeting. Meta increasingly wants advertisers to trust its systems to identify the people most likely to respond rather than manually defining every characteristic of the audience themselves.
According to Mari, the results advertisers are seeing aren't universally consistent. Some report improvements in their return on ad spend, while others find the automated targeting has attracted the wrong people. That doesn't mean businesses should avoid these tools, but it does mean we shouldn't simply switch everything to automatic and stop paying attention.
Test them. Start with budgets you're comfortable experimenting with and watch the results carefully. If you're investing real money into advertising, AI shouldn't become an excuse to stop monitoring where that money is going.
Your conversions may tell you more than the targeting does
One of the frustrations with increasingly automated advertising is that marketers don't always get to see exactly why the system has selected particular people. Meta knows considerably more about the processes behind its algorithms than it is ever likely to share with advertisers.
That means your own first-party data becomes even more valuable.
Look at who is actually signing up to your email list, submitting enquiries, booking appointments or buying your products. You may discover that the AI has identified a group of potential customers you hadn't previously considered. Equally, you may discover that it is generating plenty of activity but very little meaningful business.
This is another reason why relying purely on reach, impressions or clicks can be misleading. The real question is what happens after somebody sees or clicks the ad. If the people Meta is finding are converting into customers, that's useful information. If they aren't, you need to know that too.
Don't let AI make you sound like everybody else
There is another problem we're beginning to see as more businesses use the same generative AI tools: everything can start sounding remarkably similar.
Certain words and phrases crop up repeatedly. Suddenly everybody is "unlocking" something, "leveraging" something else or discovering the "secret" to whatever they're selling. If businesses simply accept the first version an AI tool produces, we risk creating an internet filled with perfectly competent content that all sounds as though it was written by the same person.
Your brand guidelines therefore matter more, not less, in an AI-powered marketing environment. Your tone of voice, language, opinions, expertise and the way you communicate with customers should influence anything produced with AI.
The same applies to advertising. If Meta suggests several variations of ad copy, don't assume the system knows your brand better than you do. Use the suggestions as a starting point, rewrite them and make sure the finished version sounds like something your business would actually say.
It's also worth keeping an eye on what your competitors are doing. Meta's Ad Library allows you to see advertising being run by other businesses, which can help you recognise when everybody in your sector is beginning to use the same language, formats or approaches. Sometimes knowing what everyone else is doing is exactly what you need to work out how to do something differently.
Automation can help customer service, but people still need people
AI-powered chatbots are another area likely to develop considerably. Businesses can already use automated responses through social platforms and third-party tools, and these systems are becoming increasingly sophisticated.
The potential is obvious. Imagine being able to train an AI assistant using your FAQs, products, services and customer information so that it can answer common questions, help people make buying decisions and provide support before and after a purchase. For a small business, that could save a significant amount of time.
The danger comes when automation becomes a barrier rather than a help.
We've all experienced customer service systems that send us around in circles when all we want to do is speak to an actual person. Mari's point was that, regardless of how sophisticated these systems become, customers should have a route to human support when they need it.
Automate the repetitive questions if it makes sense for your business, but don't forget you're communicating with people. Relationship marketing doesn't suddenly become irrelevant because technology has become better at answering questions.
Will AI replace marketing jobs?
It's impossible to have a conversation about AI without eventually getting to this question. If AI can create content, answer customers, produce advertising, analyse performance and eventually make increasingly sophisticated strategic recommendations, why would businesses continue hiring marketers?
There will undoubtedly be disruption, particularly in areas where the work is highly repetitive or where businesses believe they can achieve a good enough result using automation. Content creation is one area where this is already becoming particularly noticeable, and AI systems are likely to become much more capable over the next few years.
However, there is a significant difference between producing marketing material and understanding a business, its customers, its positioning and the wider context in which that marketing appears.
The role of the marketer may change, but judgement, strategy, creativity and the ability to understand people remain incredibly important. If anything, marketers may increasingly need to understand how to direct, evaluate and improve what AI produces rather than simply creating everything manually.
Don't automate the relationships out of social media
Perhaps the most important message from our conversation with Mari was also one of the simplest: treat people like people.
Social media was built around human connection. Businesses have benefited enormously from being able to talk directly with customers, build communities and develop relationships that would have been considerably more difficult before these platforms existed.
AI can make some of that easier. It can save time, generate ideas, help us analyse information, improve advertising and handle repetitive tasks. What it shouldn't do is remove every trace of the person behind the business.
There is a temptation, particularly when you're busy, to automate as much as possible. If AI can create the post, generate the image, schedule it, respond to the comments and answer the customer, it may eventually be possible to have an entire social media presence running with very little human involvement.
But at that point, we have to ask what exactly is still social about social media.
Use AI, but keep questioning it
There is no reason for small businesses to avoid AI. Used properly, it can provide access to capabilities that previously required considerably more time, money or resources. It can help you get past the blank page, find new ways to use existing content, test advertising ideas and become more efficient.
What we shouldn't do is assume that because the technology is impressive, it's automatically right.
Test it, question it and edit what it gives you. Monitor the results when you allow advertising platforms to make decisions on your behalf. Be transparent where AI-generated media could mislead people, and make sure your brand voice doesn't disappear into a sea of AI-generated sameness.
Most importantly, remember who all of this marketing is ultimately for. Your customers aren't algorithms, impressions or data points. They're people.
And however sophisticated AI becomes, that is something worth keeping at the centre of your marketing strategy.
