Trending Articles
Trending Topics
Trending
Rachel Fairley might be best described as a fixer. Known for her work in integrating brand and marketing to drive growth, she even wrote the bestselling book Rebrand Right on the subject. Her career has spanned over 35 business transformations across industries and countries, advising market leaders on everything from finding new revenue to improving employee effectiveness.
Thanks to her experience guiding brands through transformation, Rachel approaches AI from a pragmatic lens. She sees the value AI adds when it comes to “fixing leaks in the experience to convert more buyers,” but she’s equally quick to point out that “whatever method you use, nothing is perfect and all is useful”. For Rachel, AI should complement human intuition rather than replace it, especially when building trust and authenticity in brand marketing.
Rachel’s view of ROI is similarly rooted in the bigger picture. Drawing on her experience steering decision making, she stresses that measuring AI’s impact alone misses the point. For Rachel, “the only ROI that matters is revenue”.
AI is a deeply powerful tool, but only effective when paired with cohesive strategy. Her approach mirrors the “fixer” mentality she’s built her career on, focusing less on the novelty of the technology and more on how it can contribute to long-term brand trust and growth.
Building on her deep expertise in brand transformation, Rachel understands that data and insights are only as valuable as the changes that they can inspire. She knows first-hand how emerging technologies can fit into understanding customers and business growth.
It’s thanks to this experience that she’s in a better position than any to answer our first question of just how effective AI can be when it comes to predicting customer behaviour.
It really depends on the ‘exam question’ being answered.
For fixing leaks in the experience to convert more buyers, increase the pace of their journey, better cross or up-sell, or reduce churn, then it is all about understanding the existing patterns of behaviour. Looking at insights from company data and how that is understood by both the humans working with these buyers and AI’s interpretation is all invaluable. The challenge is what data the predictions are based on. I always work back from point of sale (be it new sale, upsell or resell) to awareness, because fixing those nearest to the sale is the priority (think of how much work you have done to get the buyer to that point!).
When I’m focused on increasing market share (so pretty much always!) or extending into new markets this means wooing new buyers, who may or may not behave like existing buyers. There I find research works best – qual, quant, synthetic.
Synthetic research can be helpful even though it won’t give you the surprises and oddities found in human research that can inspire really different ideas, or the speech or physical expressions of real people. Michael Mace, VP of the Center for Human Insight, explains why this matters: “The quirkiness, diversity of answers, and variety in phrasing that you get from human interviews feed your intuition. In contrast, the synthetic participants feel averaged out, like a stone sculpture that’s been sandblasted until you can only see its basic outline.”
It is worth understanding what data the intelligence is using to inform its thinking. Is there a bias in the data? Caroline Criado Perez explains in Invisible Women, the sex and gender data gap may mean the answers are based on the average “default male” who may not be your buyer.
Whereas when talking to sales and leadership you need to watch out for ‘recency bias’. Do backwards research so you are looking for explanations for what you see in quant, rather than on a fishing expedition.
Macro-economic and political events can change how people behave. This can mean past performance is not an indicator of future performance, so always take the predictions with a pinch of salt.
Whatever method you use, nothing is perfect, and all is useful. It is worth seeing over time which becomes more reliable for your organisation. Humans are both predictable and unpredictable – aren’t we fabulous! But the better the data, the more you can see the shape of behaviour and find where the experience is less effective at driving the actions your business needs to grow revenue. Yes, the tech helps, but use your brain too and speak to real buyers. Just don’t forget to look for the quirks, not just the ‘typical’.
[Rachel points readers to the book Invisible Women by C. Criado Perez and a blog by M. Mace, ‘Can AI replace discovery interviews. A competitive comparison’, Center for Human Insight.]
For years the criticism of business-to-business marketing has been that it isn’t authentic unless it combines emotion and logic in its marketing. And yes, many b2b brands have been repressing their emotions. Now we’re worrying whether content authored by people is more or less authentic than content made by AI. Whatever the buzzwords, can we park them?
Some say we live in a post-truth age. Edelman’s trust barometer shows trust in institutions is down. It’s a sign of how unsettled our world is politically and economically. If you want your brand to thrive it has to be trustworthy. Authenticity is about building trust. Because who buys from a brand they don’t trust?
If you want your brand to be authentic, it has to be clear and cohesive about why it exists, what it does, how it does it and be easily recognisable. The experience has to be easy and cohesive.
The buyer is human, even if they are buying for their business. And people are complex, unpredictable, quirky. You have to orient your business to help them find what they are looking for when they come shopping, to achieve the benefits they are seeking. Market oriented, not founder or product-oriented or organisation-chart oriented.
You need your brand to make a connection with the buyer emotionally. This builds your brand associations in their minds, making you easy to mind in a buying situation, earning you trust and giving them confidence in choosing your brand.
Can you do that using tech? Of course. But if you pollute their world with an overwhelm of average content, good luck to you. Avoid making derivative works (where you can clearly see the lineage) and focus on transformative (new) works. Random acts of marketing can make marketers feel like they’ve delivered even if it reduces the impact of the brand overall. If your competitors zig into that route that might give you an opportunity to zag and be fresh and cut through the noise. Make sure your content is cohesive, relevant, easy and different.
Everything your brand does builds or destroys the trust it has earned.
“37% of chief marketing and communication officers believe AI poses a brand risk. More broadly, 38% of all respondents admit they are unclear about the ethical implications of using AI – a figure that rises to 41% among B2C professionals” reports Editor Elizabeth Howlett in PRmoment.
Generative AI’s widespread use has introduced new uncertainties. As with all technology advancements, you have to take time to understand the limits and implications and make changes to best practice and governance.
Otherwise, it can undermine trust. “Data theft, labor exploitation and outsized environmental impact” are three ethical considerations Emily M.Bender identified. Gleaning intelligence from biased or incomplete data can influence the wrong behaviour or decisions. Mis-reading the audience’s expectations of how personalised the interactions with the brand are (‘you know too much about me’ versus ‘why don’t you remember me’) will be a careful balance. Mandating AI without training can alienate staff. Promising no job cuts because of AI and then doing them anyway can also undermine the brand. Agentic AI requires, at a minimum privacy and data barriers be dramatically changed.
This is all about reputational risk. Chief Communications Officers know that “…one poorly timed or poorly worded message can spark backlash, erode trust or derail broader strategic goals” finds MikeWorldWide in their ‘Confidence Amidst Chaos’ research. The many communications and marketing leaders I talk to say they feel like they are walking on quicksand. Inertia isn’t an option, so making well-considered decisions, adjusting governance and ways of working, is imperative.
[Again, Rachel points to sources: Emily M Bender, E. Howlett, ‘A third of in-house comms and marketing pros feel overwhelmed, research finds’, PR Moment, and ‘Confidence Amidst Chaos’, MikeWorldWide]
You can absolutely measure the impact of specific tactics in a marketing programme, whatever powers them. Every tactic should have a purpose, and monitoring if it is achieving that is a necessity.
But you do need to pull out of the myriad micro performance indicators and take a macro view. The only ROI that matters is revenue. And for this, you need a strong brand, an integrated brand, marketing and experience strategies.
Strong brands build trust. People buy from brands that come easily to mind, that they know of, recognise and are familiar with. They don’t buy from brands that they can’t remember, don’t recognise, don’t know about or aren’t convinced by. And it is exactly the same for talent. It’s so attractive and sticky working for a strong brand, which means it is high in cohesion, relevance, ease and difference (CRED).
Strong brands are efficient. Every tactic, interaction, and experience builds on the associations already in the person’s mind. Making performance/growth/demand marketing (pick your favourite term!) perform better.
This is never down to one tactic. Or whether that tactic was AI or human-made. And if you choose between investing in brand or demand, you won’t win market share. You may be able to pull apart the tactics that were AI-powered from non-AI-powered, but if your buyer or talent is experiencing them all, then how can you ever know what the true impact on revenue is?
I get that we’re all suddenly excited by artificial intelligence. With blockchain, cloud and machine learning, it has been a great enhancer of our tech for a long time now, making our marketing work more efficient and scalable. Production tools with artificial intelligence and machine learning built in are brilliant for scaling, making all the variants needed for an optimised execution. But if you are being asked to prove AI can do a better job than humans, it might be worth considering the implications for you and your team and the buyers who experience your marketing.
[Sources: Brand Finance 2024, ‘Findings. Distinctiveness by numbers’, JonesKnowlesRichie]
Do use your brain and do optimise your brand for LLMs!
MIT published a study of what happens to your brain if you write essays over four months using LLMs, versus search, versus brain. N. Kosmyna and colleagues used electroencephalography (EEG) to assess cognitive load during the writing, analysed essays using NLP, and scored them with the help from human teachers and an AI judge. To summarise – brain-only users best brain connectivity, search-users moderate and LLM-users weakest. They reported “LLM users consistently underperformed at neural, linguistic, and behavioral levels.”
Gavin Harvey in The Nuron commented, “The most concerning part was this: 83% couldn’t remember a single sentence from the essays they’d just written, compared to 11% of those in the brain group. Even scarier, most people had no idea their thinking was being affected. The convenience feels seamless, but the cognitive trade-offs were invisible until researchers measured what was actually happening upstairs.” He added, “The MIT researchers coined a term for this: cognitive debt. You’re essentially borrowing against future cognitive capacity for short-term convenience.”
Please don’t outsource your thinking. Use AI for refinement. Be wary of its data-bias, hallucinations and vanilla-ness. It is a fab tool, yes. Keep building your own expertise and find people who are better than you at what they do and work with them.
Do optimise your brand for LLMs. For this, you need to be everywhere with quality content so that next time the AI companies train on fresh data, your brand’s content is included. AI overviews are effectively AI-powered summaries of the top traditional search results, so your SEO team (and communications who know all about earned and owned) with quality content is your answer. Ask AI about the industry and your brand, and pay attention to the sources it cites, for that is where you need to show up with quality content. Look at your website data to understand which pages of your website the AI finds most useful.
[Sources: Grant Harvey, ‘Your Brain on ChatGPT is Accumulating “Cognitive Deb”… Or is it “Decline”??’, The Nuron, 17 June 2025; N. Kosmyna et al’s study, ‘Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task’ published by Cornell University on 10 June 2025]
