Marketing In The Age Of AI: Dos And Don’ts To Maintain Brand Authenticity

There’s no escaping the surge of AI across all aspects of our lives. In 2025, 32.7% of EU citizens aged 16-74 used generative AI, rising to 64% for ages 16-24 (Eurostat).

For some, AI is even threatening their livelihoods. Anthropic CEO, Dario Amodei, has predicted AI could eliminate half of all entry-level management jobs within five years. A sobering number.

So what does this mean for marketing?

Whether you work as a marketer in a big corporation, or you’re a founder building your own brand, AI has laid claim to many aspects of the marketing role. From brand strategy, to visual and video creation, AI is the shiny new tool that helps get the job done faster, but not always better. 

In this article, I will lay out what I have seen so far and the Dos and Don’ts of using AI in your marketing.

1.  Research

It’s so tempting to cut down on all that laborious work. And yes, AI can indeed help you become much much more efficient here. It can also see patterns that may not be immediately obvious to the human eye.

But AI cannot go to a store and see what’s happening on shelf, it cannot see how consumers are behaving and interacting with the product. It cannot talk to your ideal clients and hear the words they actually use.

So with this in mind, here’s how you can use AI efficiently in your research, without losing the human touch.

• DO use AI to help search through a vast quantity of information available. Instead of spending hours searching yourself, accurately prompt your AI tool to do the research for you.
• DON’T rely on AI to find all the information available. AI can easily miss the subtleties of the right competitors and will not be able to anticipate any future moves they may make. Double check the facts, go to stores, do your own SWOT. For example, I’ve seen AI miss a key competitor when I was doing research, simply because the name was similar to a brand it had already spotted (it assumed it was the same one!). When I questioned it, it apologised profusely for the confusion. Nothing replaces human knowledge.
• DO use AI to help you think through and structure interviews and client research.
• DON’T rely on AI to give you the answers. You need to observe and talk to your clients yourself. You need to hear the words they use, understand the nuances and feel their emotions.

2.  Strategy

Again, it’s tempting to not do the real work here and simply ask AI to write up your positioning, differentiation and messaging strategy for you. And yes, AI can be a great brainstorming buddy (I use it all the time for this!).

But it can’t replace your human brain. We can see possibilities that AI can’t, simply because AI can only see what already exists. We can also feel what works and what doesn’t, in the way only a living human being and our sense of intuition, can.

In terms of strategy:

• DO use AI as your thinking partner to help analyse key competitors’ positioning and messaging, brainstorm ideas and share pros and cons of different approaches.
• DON’T rely on AI to write your brand strategy for you. AI can’t see beyond what exists. If you're launching something the market hasn't seen yet, a new ingredient, a new movement, a new technology or aesthetic, AI has no reference point. How can it then help find a positioning it can’t see? Only your human mind can imagine possibilities beyond what exists today.

3.  Content Creation

This is the area where most AI is used in marketing today. But it is also, in my view, the most controversial. Legal frameworks have yet to keep up with the boundaries around AI-generated marketing content such as usage rights. Which means that AI-generated content is solely an ethical issue. And ethics are intricately linked to brand values and purpose.

Let’s look at the beauty industry: Dove has been advocating for Real Beauty for over two decades now. In one of its latest campaigns, Dove showcases how AI interprets “the most beautiful woman in the world”. Fake, overdone, computer-generated. Whilst Dove commits to never using AI models in its own advertising, AI-generated influencers now have hundreds of thousands of followers. Check out Aitana, an AI model who is making hundreds of thousands of Euros through sponsorship by sports and beauty brands:

The risk is real that influencable, younger generations will not know the difference between what’s real and what is not and will aspire to a fake reality of both beauty and lifestyle.

The fashion industry is also experiencing its own dose of AI backlash. From a UK GAP ad being banned for “abnormally thin legs altered digitally” to J Crew’s latest campaign under fire for alleged use of AI models that simulate diversity without using real people (mismatched lines on the jumper and a missing finger gave the game away!).

Conversely, the 2025 Christmas ad by French retailer Intermarché, "Le mal aimé" ("Unloved"), went viral by combining high-quality, handcrafted 3D animation with an emotional storyline about a lonely vegetarian wolf, deliberately contrasting with AI-generated content to promote themes of belonging.

AI technology is only growing more sophisticated and it’s becoming harder to differentiate AI from reality. So amidst this craziness, how do you keep creating content, whilst maintaining your brand values and authenticity?

• DO use the snazzy new features available to help create your visual and video content faster. Product imagery is typically an area in which AI can help generate many different options, faster and cheaper than a shoot.
• DON’T ask AI to create model shots or lifestyle content for you. And if you do use AI-generated imagery, don’t use it without disclosing it. Beyond the ethics, there's a growing regulatory and consumer trust issue here, as we saw in the J Crew example earlier.
• DO use AI to generate ideas and help polish your written content.
• DON’T copy any output directly. AI’s tone of voice is easily recognisable and sounds fake to the human brain (short, punchy, and clichéd sentences, “it’s not x, it’s y” framing, series of three, em dashes, etc.). I’ve tried using AI to write my own content. It’s definitely faster, but it just doesn’t sound like me (and I’m pretty sure my audience senses it too). Here’s a recent example from an article I couldn’t bring myself to publish:

Sustainability doesn’t change behaviour.
Desire does.

But here’s the harder truth:

Even when you understand this… you may still be playing too small.

Not in effort — but in how you’re thinking about the game.

Whilst I agree with the message (I shared my ideas with ChatGPT so it could write the article), I would simply never write like that myself (“here’s the harder truth…” really??).

• DO use AI for repurposing content. It’s a brilliant tool to help you take a longer piece of content (eg. a long-form article like this one) and suggest how to break up the ideas into several posts for social media or newsletters. It can also alter visuals (different angles, B&W vs color, etc) to keep things looking fresh on your feed without creating new material all the time. This keeps your original brand identity as the source material, whilst speeding up your daily content output.
• DON’T simply copy and paste these shorter content pieces. You still need to do due diligence in checking the quality and ensuring they look and sound authentic to your brand.

4.  Community Building

This is going to sound obvious, but to build a community, you need to create real, authentic connections. Not ones that are managed purely by bots (duh!).

With that in mind:

• DO use AI to build a simple tone-of-voice guide, capturing your brand values, character, words you consistently use or never use, so that whoever manages the community is clear on the brand voice. You can also use AI to monitor what your community is already saying about your brand.
• DON’T use AI to respond to your community. Comments, DMs and reviews are where brand relationships are built or broken. Automated or AI-generated responses are increasingly detectable and feel deeply transactional. Invest in a human to create these connections.

Lush is an interesting example here. Known for its strong values towards people and planet, the brand will always prioritise ethics over convenience. In terms of customer care, they use AI to handle routine, high-volume tickets (freeing humans up), but their human team still manages real relationship-building. As their customer care manager put it: "we really do care about people, the environment and animals, so we won't compromise on how we treat our customers, staff or suppliers, just to make things cheaper, easier, or quicker.”

The Recap:

The simplest way to summarize how to leverage AI without sounding fake and inauthentic in your marketing is this:

• Use AI to help with your research, as a brainstorming partner, and for support with your content creation.
• Don’t rely on AI to have all the answers or to create direct output for your brand.

AI is improving at such an exponential rate that what is true now, won’t be in just a few months. But one thing will remain true: nothing can replace our humanity, our values and our sense of right and wrong. Our human brains have an innate sense in terms of authenticity, emotion and connection, even if we can’t explain it.

So my best advice is this: if something doesn’t feel right, don’t use it.

In a world of AI-generated everything, your most powerful differentiator is your humanity. If you want help putting that at the heart of your brand, you know where I am. :-)

Full Disclosure: How I Used AI To Help Me With This Article:

• I first wrote a full, messy draft. I didn’t pay attention to typos or links to examples I was writing about, I literally “bashed it out”.
• I shared it with Claude (I recently switched from ChatGPT to Claude for ethical reasons). I told Claude who my ideal clients are and what my POV on marketing is and asked for its feedback. I explicitly said “don’t rewrite anything, I want it to stay in my voice”.
• Claude shared basic feedback (grammar, typos) and suggestions by section for me to implement.
• I chose to implement some of the suggestions – not all. I integrated them into the article myself.
• I shared the reworked document with Claude for a final check and corrected remaining typos.
• I will now ask Claude to make some suggestions on how I could break up the ideas in this article to create several posts for Linked In.