AI is everywhere, but does that mean you should be using it? 

Artificial intelligence software, like ChatGPT and muse.ai, are making the rounds across news and social media sites. And though it’s pretty funny to ask a computer to write a heartfelt, detailed critique of Shrek 2 (trust me, I tried), it has sparked conversations.

Is AI affecting customer relationships and brand authenticity? Will we still have jobs in the near future? Can AI help one person with everything? (Spoiler. Nope!)

So, is AI friend or foe? Let’s start from the top. 

What is AI? 

AI (artificial intelligence) is technology developed to perform tasks that, in most cases, needs human thought and interaction. Visual perception, recognising language and communication, and decision-making, for example. 

We haven’t reached Blade Runner or movie-like levels of AI…yet… however, in recent years, it has increased significantly in popularity. Why though? 

As computer power and processing improve, AI improves alongside it. Unlike us, data can be stored and never forgotten on computers. With more data to play with, AI can piece together more human-like patterns and start performing all the tasks mentioned above. 

It’s basically a robot elephant that never forgets.

Why is this relevant to marketing and advertising? 

In digital marketing, what’s essential to make sure you understand audience behaviour and patterns? Data. And who better to ask for help than a machine that’s literally got all the data? 

Take Google, for example, which harnesses AI machine learning by modelling data from your ads and campaigns – which you can then make sense of with Google Analytics 4 to forecast trends and provide actionable insights. 

AI can observe patterns of behaviour – time of day, location etc. – then predict and recommend when and where you should place your ads. 

This is just the tip of the AI-ceberg. If you link all your Google properties together – like Google Ads and Campaign Manager – it can combine these data points to train machine-learning-driven bid strategies. This fuels your data-driven attribution models for a better understanding of your media mix.

These data-driven machines will ultimately improve customer experiences as they will report on what’s working and recommend relevant desires – increasing ROI for your campaigns, decrease Ad Spend, and time needed to gather data.

However, and this is a big however, AI is still not at the level of understanding data as well as agencies who live and breathe it. Much like creative, ads need a human touch and level of management to be utilised properly. Think of AI as a tool businesses can use to streamline certain aspects, but should not be left alone to monitor data by itself. 

For the people? 

Data analysis is one thing, but as we continue to type away and chat online, AI is taking notes. 

Recently, instead of teaching AI how to talk the talk, we’ve told it to walk the walk, asking it to learn language for itself. By throwing an almost infinite amount of written words, sentences and articles at it, AI has evolved to predict what to say and says it pretty coherently. Researchers refer to this as a language model. 

Because of this, chatbots like ChatGPT can be super useful in conducting basic keyword research on specific topics and surfing the internet to recommend great starting points in blog writing or ideas for a presentation. 

But don’t rely on AI to write all the copy for you. Its monotone and repeated descriptions of certain topics can make blog posts dull and sound automated, not to mention its similar responses to topics can cause a negative impact on SEO. Also, my critique of Shrek 2 has much more detail. 

Can A–I help you?  

Have you ever been on a website and a little box in the corner pops up saying, “Hi, my name is Jenny! Can I help you with anything?”. Aw, that’s nice, sounds like somewhere, Jenny is taking time out of her busy schedule to ask how I’m doing.

Sorry, Jenny doesn’t exist. 

This is an AI chatbot, a software used to help answer a customer’s question on a site. Marketers will implement several keywords for the chatbot to identify (also known as a playbook) as simple as “Yes” or “No” and direct them to the solution. 

For example: 

AI – “How can I help?”

User – “I want to return my shoes.” 

Result – The AI detects the keywords and can explain or direct users to the item return page or recommend exchanges for different shoes. 

As language models become richer in detail, chatbots can detect more keywords and customer queries and respond to niche questions. As a result, this gives marketers more time to focus on other important tasks, as chatbots can run continuously 24/7. 

Chatbots are great if you’re handling multiple websites or social media accounts and need time to keep track of other factors. 

We’d always recommend that you employ someone who’s available to chat with customers, though. Even if AI can answer questions, it will lack that spark from your brand’s Tone of Voice, the one you’ve spent years building up, and potentially cost you losing consumers so close to the end of the customer journey.  

AI gets bigger and better 

If you have a phobia of android overlords taking over, you might want to cover your ears. Artificial intelligence is expected to generate roughly £30 billion in revenue by 2025, and brands aren’t backing down in taking advantage of its potential. 

Netflix is no exception. As one of the biggest VOD platforms worldwide, you can imagine it’s a daunting task for marketers to cater to the needs of each individual user. UX is a major factor for customers, hence why Netflix use AI to make it all the easier. And here’s how…

Everything from personalised show and film recommendations to optimised streaming quality based on internet speed is all monitored and managed by Netflix’s machine learning. And the outcomes speak for themselves – over 80% of the shows people watch on Netflix are through its AI recommendations. 

So far, sounds like AI is A-OK, right? Well, let’s not get too carried away! AI is great, but it isn’t perfect; there’s a lot you need to be aware of before implementing it into strategies and tactics.

AI has a time and place, and knowing when and where to use it is our (and should be your) mindset.

Creative and AI?  

One of the biggest concerns is from the perspective of creativity. Is AI taking creative work away from creatives? The answer is no. Not where creative matters anyway.

In fact, creatives are utilising AI to its full potential. In an interview with The Drum, Invisible North’s Geoff Renaud states: 

“By using AI to automate outputs and shrink hours needed for tedious tasks, deliverables will get done faster, more efficiently and minimize costs for agencies and their clients. While human editing and oversight for most of these tasks will be required in the near-term, data is the prime currency for these tools and machine learning will rapidly compound results as usage increases and these new tools improve at an incredible pace.”

There lies the calming clarification – AI isn’t here to replace, rather reduce the tediousness of less creative tasks. 

If you ask ChatGPT to write copy, it will do so, but you’ll know an AI’s written it. The same goes for visual creative – ask AI to draw a duck; it’s going to draw that duck the same way, again and again, unless you start tampering with the conditions. Eventually, AI creativity runs dry.

For now, AI can’t add flair or a distinct tone of voice, nor can it give more than one opinion; what’s fact is a fact for Chat. 

Ethics and AI

On the flip side, AI raises ethical concerns when discussing data and audience targeting. Why is this? 

Well, back on the topic of data and AI’s objective outlook on stats, it’s very easy for machine learning to mistake numbers and percentages for identities; people are simply more than quantitive lists on a sheet of paper. 

This can lead to subject bias, where AI takes “stats” regarding personal information – gender, age, race, location etc. – and unintentionally delivers your ads to people who share these factors even if they don’t desire the product you advertise. 

This is a massive red flag; you want to deliver campaigns based on targeted audiences, not people who fall under a certain “demographic” – that’s borderline discriminatory. 

And with only 48% of customers trusting companies to use AI ethically and 65% concerned about the unethical use of AI, you don’t want your brand getting called out for something Mr Robot worked on.

So, what do we recommend? 

AI for creative is another tool to use, not a crutch. 

AI offers a wide range of rich data, keywords and topic starters to get you in the zone. The same goes for design and illustration; it’s a great catalyst to get creativity flowing. 

However, it’s not – and should never be – the end product. If you rely too heavily on machine learning, eventually, someones going to catch on. Creative still requires that “human touch” and emotional connection. After all, we know ourselves better than the machine does. 

Just like next door’s dog, don’t leave AI on its own. 

AI is far from perfect. Chances are, it will eventually make a mistake and pick up on rubbish data, awful ideas or discriminatory ideologies.  

Have someone monitor your AI processes; no “copy and pasting into the sales deck, job done!”. Make sure the data collected resonates with your audience personas and aims to deliver ads for the people who want to see them. 

Be Transparent 

There’s no shame in using AI; everyone seems to be. Netflix and Google openly admit that they use machine learning to gather data, streamline processes and hone in on audience preferences. 

So why shouldn’t you? Be open about your privacy and data policies. This rule has always been a given, but with AI now in the mix, it’s always good to remind customers of what data is kept and optimised to help ads find their way to them. 

Most importantly, yes, you still need an agency! 

AI doesn’t always lead to the best results. We saw Performance Max make its way onto the scene, among other automated ad products like Meta Advantage. But there are still question marks over its effectiveness in boosting conversions. 

Sam Dadd, Our Junior Paid Media Executive, had this to say about machine-learning systems: 

“Having recently tested Performance Max (P-Max), we have been happy with some aspects of performance achieved, mainly with conversion-driven campaign objectives for e-commerce. However, lack of control and transparency provided by Google on P-Max campaigns limits the ability to learn, make data-informed decisions, and scale performance.

P-Max campaigns will optimise towards ‘quick and easy’ conversions, which means a strong focus on brand terms in paid search with little consideration of incremental impact. This may cause higher CPCs or CPAs overall, whilst potentially cannibalising other aspects of your marketing strategy.

It’s important to remember that you’re giving ad platforms free rein to spend your budgets with very little visibility on where, how, and to whom your brand has appeared.”

In summary, you’ll need an agency of experts in their field to craft personalised media strategies unique to your business goals, plan and execute campaigns that stand out from the noise, and deliver results with agility and maximum ROI in mind through careful optimisation.

We can help!

Remember, to um is to be human, and to take that extra moment and think about what needs doing, whether creative, paid or in marketing, will make all the difference. But for now, we can Domo Arigato Mr Roboto for letting us get on with the motive that matters most – making the customer happy. 

Want to implement AI machine learning into your marketing strategies? Eager to produce next-level creative with that “human spice” for an upcoming campaign? Or asking, “wtf is Google Analytics 4, and how do I use it?” – Contact our team of experts, and we’ll help create strategies in paid and creative that feel right on brand.