Marketing, AI and the Future of Marketing
Marketing is understanding people. AI is changing how people behave.
Nifemi Aikomo
Director of Strategy, ImpressionsON
Nifemi Aikomo is Director of Strategy at ImpressionsON, leading social listening, consumer insight and campaign strategy. In almost ten years in marketing, he has worked with Union Bank, Heineken, Paystack, FSDH and AVA Energy. He also keeps a range of creative and entrepreneurial pursuits, and holds an MSc in International Business with Distinction from the University of Reading.
Chapters
01
“
Where I’m starting from
02
61%
Marketing is understanding people, and people are changing
03
“
Audiences are not static beings
04
99%
How I think about AI
05
80%
AI in Nigeria is here to stay
06
48%
How AI will change marketing
07
3.6x
Six shifts I see coming
08
“
So what should marketing do?
09
“
To close
In this report
Chapter 01
Where I’m starting from
When I look at what AI is doing (and this might just be me in my own bubble), I see the same thing.
It’s taking work that only a few people could do and handing it to everyone.
I’ve never put AI in my bio, and I’ve never tried to position myself as an AI specialist. But I’ve been intrigued by it for a while now, and not even mainly within the space I work in. I’m intrigued by what it makes possible.
A few years ago, if you wanted a landing page, you needed a UI designer, a web developer, someone to write the copy, and a budget to pay all three. Today, with a $20-a-month subscription to Lovable, you can get a landing page done in an afternoon. It might not be great. It might look like ten other landing pages on the internet. But you still get a landing page, and for most people that would not have been possible before.
A few years ago, if you wanted a landing page, you needed a UI designer, a web developer, someone to write the copy, and a budget to pay all three. Today, with a $20-a-month subscription to Lovable, you can get a landing page done in an afternoon. It might not be great. It might look like ten other landing pages on the internet. But you still get a landing page, and for most people that would not have been possible before.
What a landing page takes
A few years ago
04
- A UI designer
- A web developer
- Someone to write the copy
- A budget to pay all three
today
02
A $20-a-month subscription to Lovable
An afternoon
One of the first campaigns I ever worked on in marketing, on the strategy side, was Union Bank’s “Don’t be an Uncle Thomas” in 2018. Union Bank had just turned 100 and rebranded the year before, moving away from “Big, Strong, Reliable” to a simpler, smarter bank, trying to shake off the image of slow, old-school banking. The campaign played on “Thomas,” our Nigerian expression for the doubter (from the disciple who wouldn’t believe until he saw), and asked Nigerians not to let disappointment turn them cynical about their own country. When we were thinking about the brand archetype, we went with the democratiser. The brand should take what was only available to the 1% and put it in the hands of the 99%.
Union Bank, “Don’t be an Uncle Thomas” (2018): “Nigeria we hail.”
When I look at what AI is doing (and this might just be me in my own bubble), I see the same thing. It’s taking work that only a few people could do and handing it to everyone. And it will only get better. Designs will get cleaner, logos will get better, and people will get access to information they never had a way into before.
Two ways I’m intrigued
This is not me saying I’ve become an AI expert or that I know everything about it. Like most people, I’m intrigued from a couple of angles. The first is as a person. What could this do for my life? What data am I sharing every time I type something in? What are these tools learning about me? You see all the doom-scrolling content about what AI will do to us, and I’m just as curious about what it can do for us.
The second hadn’t really crossed my mind until recently, when I attended the Meltwater Summit. It was how AI matters to the field I’ve given almost a decade of my life to, which is marketing. I’d never sat down to think about it properly. At best, I thought about how it could speed up my research. I had never thought about it from the fundamentals of what marketing actually is. After the summit, I started to see how AI, over the next few years, is going to change marketing at its root.
The second hadn’t really crossed my mind until recently, when I attended the Meltwater Summit. It was how AI matters to the field I’ve given almost a decade of my life to, which is marketing. I’d never sat down to think about it properly. At best, I thought about how it could speed up my research. I had never thought about it from the fundamentals of what marketing actually is. After the summit, I started to see how AI, over the next few years, is going to change marketing at its root.
Chapter 02
Marketing is understanding people, and people are changing
61%
A year ago, when people in the US compared iPhones, 63% did it on Apple’s comparison page and 37% did it with an AI tool. Twelve months later, it has flipped: 61% now compare with an AI and only 39% use Apple’s own page.
At the core of marketing is understanding people: who you’re selling to, how they behave, how they react to things, and how you can reach them. I know that’s as simplified as it gets, and my apologies to every esteemed marketing professional reading this. But with the knowledge I have, that’s what I believe marketing is. You understand human behaviour, and you use that understanding to position a product.
What AI is doing, and this is why it matters so much for marketing, is changing how people behave.
One of the first things you learn in marketing, in school or on the job, is the funnel. Awareness, consideration, conversion, and a few more stages depending on who’s teaching it. Alongside the funnel, you learn which platforms and touchpoints reach people at each stage.
Look at that funnel today. Discovery still runs through search. In February 2024, Gartner predicted that search engine volume would drop 25% by 2026 because of AI chatbots. That didn’t happen. Gener8, which tracks the real browsing and AI use of a consented panel of over 230,000 people in the US and UK, found that over the past year Google searches per user went up 23% in the US and 16% in the UK. That was the same year monthly AI use on mobile in the US went from 24% to 43% of people. What changed is how people search. The average Google search is now 26% longer than it was a year ago, and Gener8’s read is that people have started typing into Google the way they talk to ChatGPT.
What AI is doing, and this is why it matters so much for marketing, is changing how people behave.
One of the first things you learn in marketing, in school or on the job, is the funnel. Awareness, consideration, conversion, and a few more stages depending on who’s teaching it. Alongside the funnel, you learn which platforms and touchpoints reach people at each stage.
Look at that funnel today. Discovery still runs through search. In February 2024, Gartner predicted that search engine volume would drop 25% by 2026 because of AI chatbots. That didn’t happen. Gener8, which tracks the real browsing and AI use of a consented panel of over 230,000 people in the US and UK, found that over the past year Google searches per user went up 23% in the US and 16% in the UK. That was the same year monthly AI use on mobile in the US went from 24% to 43% of people. What changed is how people search. The average Google search is now 26% longer than it was a year ago, and Gener8’s read is that people have started typing into Google the way they talk to ChatGPT.
Search went up, and got longer
Google searches per user, past year: US
+23%
Google searches per user, past year: UK
+16%
Monthly AI use on mobile, US: a year ago
24%
Monthly AI use on mobile, US: now
43%
Average Google search length vs a year ago
+26%
Source: Gener8, How Our AI Use Changed in a Year · GenAI and Clickstream data, mobile, Aug 2025 to Aug 2026
The bigger change is at the point of consideration, where people research and compare. People are no longer doing that work on their own. They’re handing it to AI. And AI is moving into the decision too.
The example that made it click for me is Apple. Apple has a page on its website where you can compare iPhones side by side, and it’s a good tool. A year ago, when people in the US compared iPhones, 63% did it on Apple’s comparison page and 37% did it with an AI tool. Twelve months later, it has flipped: 61% now compare with an AI and only 39% use Apple’s own page. This is Apple, one of the most loved brands in the world, with a tool built for exactly this job, and people would rather ask ChatGPT about its phones than ask Apple.
The example that made it click for me is Apple. Apple has a page on its website where you can compare iPhones side by side, and it’s a good tool. A year ago, when people in the US compared iPhones, 63% did it on Apple’s comparison page and 37% did it with an AI tool. Twelve months later, it has flipped: 61% now compare with an AI and only 39% use Apple’s own page. This is Apple, one of the most loved brands in the world, with a tool built for exactly this job, and people would rather ask ChatGPT about its phones than ask Apple.
When people in the US compared iPhones
A year ago: with an AI tool
37%
A year ago: on Apple’s comparison page
63%
Twelve months later: with an AI tool
61%
Twelve months later: on Apple’s comparison page
39%
Source: Gener8, What 100 Million Real AI Conversations Reveal About Your Customers · GenAI and Clickstream data, US, Aug 2025 to Aug 2026
61%
Twelve months later, it has flipped: 61% now compare with an AI and only 39% use Apple’s own page.
People are telling AI what they like and what they want. In August alone, 35% of UK AI users typed something with a buying intent into an AI tool, things like “best Nike running shoes” or “find me a cheaper alternative to the Oral-B electric toothbrush.” Among the heaviest 10% of AI users, it was 65%. Gener8 even followed one person who added Nike Air Force 1s to their basket, went to ChatGPT 24 minutes later to ask the difference between the Air Force 1, the Air Force 1 LE and the Air Force 1 Low, then went back to Nike and bought the one ChatGPT pointed them to.
One shopper, one purchase
Nike
Added Nike Air Force 1s to their basket
+24 min
Asked ChatGPT the difference between the Air Force 1, the Air Force 1 LE and the Air Force 1 Low
Nike
Went back to Nike and bought the one ChatGPT pointed them to
Source: Gener8 GenAI, Clickstream and eReceipt data · one US shopper, 2026
There’s another part of this marketers should be thinking about. Apart from God and the person themselves, an AI tool may be what holds the most knowledge about a person right now. Say I sprain my leg and ask, “I sprained my leg, what should I do?” It gives me some advice. Two weeks later I come back and say, “I want to buy running shoes.” The best version of AI would say, “Because of your sprain, I’d recommend this type of shoe.”
Most brands are nowhere near offering that at the moment people are looking for their products. That’s where I think the biggest change will come, in buying behaviour, and in how we understand an audience at all.
Most brands are nowhere near offering that at the moment people are looking for their products. That’s where I think the biggest change will come, in buying behaviour, and in how we understand an audience at all.
Chapter 03
Audiences are not static beings
Most of the time we treat the audience as fixed: “We're speaking to Nifemi, who is 37 and lives in such-and-such city.” In reality,
Nifemi is not a static being.
The way we think about audiences has to change too. Most of the time we treat the audience as fixed: “We’re speaking to Nifemi, who is 37 and lives in such-and-such city.” In reality, Nifemi is not a static being. He changes across seasons, moods, responsibilities and money. The Nifemi you’re trying to reach today has probably changed by tomorrow. One of the clearest windows into that Nifemi is what he’s asking AI.
I don’t yet know the most effective way brands will use this, and there are real privacy questions to answer first. But it’s something we have to start thinking about.
Then there’s the grouping we love to do in marketing. “We’re trying to reach Gen Z.” Who is Gen Z? It doesn’t make sense to treat everyone between 18 and 25 as one valid target for one product. Assuming everybody in that age range likes the same thing ignores the subgroups inside it, and the subgroups inside those. A 19-year-old student in Ibadan and a 24-year-old product designer in Lekki are both Gen Z, and they have almost nothing in common when it comes to what they’ll buy and why. This is where understanding audiences through AI, through what they ask and how they ask it, can help us see people more clearly.
Then there’s the grouping we love to do in marketing. “We’re trying to reach Gen Z.” Who is Gen Z? It doesn’t make sense to treat everyone between 18 and 25 as one valid target for one product. Assuming everybody in that age range likes the same thing ignores the subgroups inside it, and the subgroups inside those. A 19-year-old student in Ibadan and a 24-year-old product designer in Lekki are both Gen Z, and they have almost nothing in common when it comes to what they’ll buy and why. This is where understanding audiences through AI, through what they ask and how they ask it, can help us see people more clearly.
Chapter 04
How I think about AI
In 2015, researchers from UC Davis and the University of Iowa trained pigeons to tell the difference between cancerous and healthy breast tissue on digitised microscope slides.
99%
When the researchers pooled the answers of four birds, the flock reached 99%.
So what do I think AI is? All protocol observed, and with respect to the people who are much smarter on this than me. I only got a proper introduction to the subject recently, and these are the two ways I’ve come to understand it.
The first is a study about pigeons. In 2015, researchers from UC Davis and the University of Iowa trained pigeons to tell the difference between cancerous and healthy breast tissue on digitised microscope slides. Every time a bird pecked the right answer, it got food. Within about two weeks, individual birds went from guessing (around 50%) to almost 85% accuracy. When the researchers pooled the answers of four birds, the flock reached 99%. Pigeons, matching trained human eyes.
Then the researchers gave the pigeons a harder job: deciding whether a mass on a mammogram was cancerous. The birds did well on the images they had been trained on and no better than chance on new ones. They had memorised the pictures. They hadn’t understood anything.
The first is a study about pigeons. In 2015, researchers from UC Davis and the University of Iowa trained pigeons to tell the difference between cancerous and healthy breast tissue on digitised microscope slides. Every time a bird pecked the right answer, it got food. Within about two weeks, individual birds went from guessing (around 50%) to almost 85% accuracy. When the researchers pooled the answers of four birds, the flock reached 99%. Pigeons, matching trained human eyes.
Then the researchers gave the pigeons a harder job: deciding whether a mass on a mammogram was cancerous. The birds did well on the images they had been trained on and no better than chance on new ones. They had memorised the pictures. They hadn’t understood anything.
Pigeons reading breast tissue slides
Individual birds, guessing
~50%
Individual birds, after about two weeks
~85%
Four birds pooled
99%
New mammogram images
Chance
Source: Levenson et al., PLOS ONE (2015) · UC Davis and University of Iowa
That’s how I see AI. It is extremely good at recognising patterns and at doing the base work that used to take people days: research, sorting, summarising, spotting what repeats. What it can’t be trusted with are the things that make us human. Intuition, gut feeling, emotion, judgment. Those are what help us make better (or sometimes worse) decisions, and they stay with us.
The second way I think about it is as the biggest library any one person has ever had, with a librarian who does the reading for you. Ask it about almost anything and it pulls together what thousands of sources have said and hands you the summary in seconds. I think that’s amazing. You want to understand the conflict between the US and Iran, and in five minutes you’ve learned enough to hold a conversation about it. (Whether five minutes of knowledge is enough is a topic for another day.)
The second way I think about it is as the biggest library any one person has ever had, with a librarian who does the reading for you. Ask it about almost anything and it pulls together what thousands of sources have said and hands you the summary in seconds. I think that’s amazing. You want to understand the conflict between the US and Iran, and in five minutes you’ve learned enough to hold a conversation about it. (Whether five minutes of knowledge is enough is a topic for another day.)
I know this is the shallowest definition there is. But it’s how I think, and how I approach conversations like this one. I’m not going to pretend I’ve become an expert.
Chapter 05
AI in Nigeria is here to stay
The numbers back this up.
80%
80% have used it to explore a new business or a career change, almost double the global figure of 42%.
At the core of all of this is one thing: AI is here to stay. It isn’t going to disappear and it isn’t going to shrink. There will be laws and regulations, as there should be, but what it can do is too useful for it to stop spreading.
Think about it from a Nigerian point of view. A mum or dad on the family WhatsApp group who would normally have to beg their second son, “Abeg, help me design this flyer,” and then wait five working days, ten working days (for a flyer that has no eventual purpose, but that’s a conversation for another day), can now go on ChatGPT, Gemini or Claude and do it themselves. Just imagine the joy. They want to tell their old school friends they miss them: “Picture of this person, picture of that person, put them together, write ‘I miss you.'” Thirty seconds later, it’s done, and it’s on a WhatsApp group, on an app installed on 95% of the smartphones in Nigeria’s big cities. You think that’s going away? Impossible.
Think about it from a Nigerian point of view. A mum or dad on the family WhatsApp group who would normally have to beg their second son, “Abeg, help me design this flyer,” and then wait five working days, ten working days (for a flyer that has no eventual purpose, but that’s a conversation for another day), can now go on ChatGPT, Gemini or Claude and do it themselves. Just imagine the joy. They want to tell their old school friends they miss them: “Picture of this person, picture of that person, put them together, write ‘I miss you.'” Thirty seconds later, it’s done, and it’s on a WhatsApp group, on an app installed on 95% of the smartphones in Nigeria’s big cities. You think that’s going away? Impossible.
Business owners who couldn’t afford a website, a flyer or a logo can now make one with free tools. That’s not going anywhere, and the sooner we accept it, the better.
The numbers back this up. In Google and Ipsos’s Our Life with AI study, which surveyed about 1,000 Nigerian adults online in late 2025, 88% said they use AI chatbots. 93% use AI to learn or understand complex topics, against 74% globally. 91% use it for work. 80% have used it to explore a new business or a career change, almost double the global figure of 42%. And 80% of Nigerians are more excited than worried about AI, compared with 53% globally. And it isn’t only a young person’s thing. Across the whole study, Gen Z use AI the most (82%), but millennials aren’t far behind at 74%.
The numbers back this up. In Google and Ipsos’s Our Life with AI study, which surveyed about 1,000 Nigerian adults online in late 2025, 88% said they use AI chatbots. 93% use AI to learn or understand complex topics, against 74% globally. 91% use it for work. 80% have used it to explore a new business or a career change, almost double the global figure of 42%. And 80% of Nigerians are more excited than worried about AI, compared with 53% globally. And it isn’t only a young person’s thing. Across the whole study, Gen Z use AI the most (82%), but millennials aren’t far behind at 74%.
Nigeria against the global figure
Use AI to learn or understand complex topics: Nigeria
93%
Use AI to learn or understand complex topics: Global
74%
Used AI to explore a new business or a career change: Nigeria
80%
Used AI to explore a new business or a career change: Global
42%
More excited than worried about AI: Nigeria
80%
More excited than worried about AI: Global
53%
Source: Google and Ipsos, Our Life with AI · about 1,000 Nigerian adults online, Sept to Oct 2025, via Techpoint Africa
So how are Nigerians using it when they shop? Visa’s Stay Secure study, published in June 2026, found that 88% of Nigerian consumers have used AI to help with shopping. The top uses were checking reviews and ratings (56%), comparing prices (54%) and finding gift ideas (53%). That’s the comparison stage of the funnel, right there. 97% say AI makes shopping online faster and easier.
But look at the rest of the picture. 83% have bought something directly through social media. Only 34% would trust an AI agent to complete the checkout for them. And 51% were hit by a financial scam in the past year, more than half of those on social media.
So Nigerians are still discovering on social, they’re comparing with AI, and they’re not ready to let AI hold the wallet. The middle of the funnel is already moving. The end of it is next, and trust is what will decide how fast.
But look at the rest of the picture. 83% have bought something directly through social media. Only 34% would trust an AI agent to complete the checkout for them. And 51% were hit by a financial scam in the past year, more than half of those on social media.
So Nigerians are still discovering on social, they’re comparing with AI, and they’re not ready to let AI hold the wallet. The middle of the funnel is already moving. The end of it is next, and trust is what will decide how fast.
Nigerian consumers, AI and shopping
Say AI makes shopping online faster and easier
97%
Have used AI to help with shopping
88%
Have bought directly through social media
83%
Checking reviews and ratings
56%
Comparing prices
54%
Finding gift ideas
53%
Hit by a financial scam in the past year
51%
Would trust an AI agent to complete the checkout
34%
Source: Visa Stay Secure 2026, Nigeria · Wakefield Research, Jan to Feb 2026, via TheCable
Chapter 06
How AI will change marketing
AI seems to agree.
48%
Meltwater analysed more than 5 million citations across the major AI tools and found that 48% of everything they cite comes from news media.
All of this matters for marketing for one reason above the rest: AI is changing human behaviour. Marketers are supposed to be the experts on human behaviour. If we miss this change while it’s happening, a lot of brands will suffer for it.
The research already shows buyers going with what AI recommends. ChatGPT alone has close to a billion weekly users, and Meltwater’s 2026 Global Digital Report ranks it the fifth most visited website in the world in 2025. Meltwater’s guide on AI visibility describes the old way and the new way well. Before, if you were shopping for a family car, you’d search “best family SUV” and get a wall of links: review sites, manufacturer sites, dealerships, YouTube videos, forums. Every one of those was a chance for a brand to make its case. Now you ask an AI and get two or three options with a short reason for each. How many people keep researching after that? Meltwater’s conclusion is blunt: if your brand isn’t one of the handful AI recommends, it will slowly become invisible.
And AI is not only giving advice. In Gener8’s UK survey of 1,020 adults in September 2026, 46% said they’d be comfortable letting AI make a purchase under £50 on their behalf. Among heavy AI users, it was 79%.
The research already shows buyers going with what AI recommends. ChatGPT alone has close to a billion weekly users, and Meltwater’s 2026 Global Digital Report ranks it the fifth most visited website in the world in 2025. Meltwater’s guide on AI visibility describes the old way and the new way well. Before, if you were shopping for a family car, you’d search “best family SUV” and get a wall of links: review sites, manufacturer sites, dealerships, YouTube videos, forums. Every one of those was a chance for a brand to make its case. Now you ask an AI and get two or three options with a short reason for each. How many people keep researching after that? Meltwater’s conclusion is blunt: if your brand isn’t one of the handful AI recommends, it will slowly become invisible.
And AI is not only giving advice. In Gener8’s UK survey of 1,020 adults in September 2026, 46% said they’d be comfortable letting AI make a purchase under £50 on their behalf. Among heavy AI users, it was 79%.
Beyond AEO and GEO
The quick answer people give to this is, “Let’s do AEO and GEO” (answer engine optimisation and generative engine optimisation). And that’s a fair answer. If AI is where people make decisions, you need to be findable by AI.
But think about how we do marketing now. SEO was never the full answer to a branding problem. SEO is discoverability. The real question has always been: does your brand show up in the right place, at the right time, when it’s meant to show up? That’s the question we should be asking about AI too.
There’s a concept for this that more of us should be using: category entry points. It comes from Professor Jenni Romaniuk and the Ehrenberg-Bass Institute, and the simplest way I can explain it is this. Category entry points are the cues that make a person think of buying in your category in the first place: what’s happening, where they are, who they’re with, what they need, how they feel. The important part is that they’re about the buyer, not the brand.
Some brands have taken the time to define these moments. A fizzy drink might say, “We show up when friends gather, we show up with food, we show up at the end of a long, hot day.” That’s the kind of thinking I mean, and it’s exactly how we should think about showing up in AI. Not “how do we rank for this keyword,” but “when someone describes this moment to an AI, are we a natural part of the answer?”
But think about how we do marketing now. SEO was never the full answer to a branding problem. SEO is discoverability. The real question has always been: does your brand show up in the right place, at the right time, when it’s meant to show up? That’s the question we should be asking about AI too.
There’s a concept for this that more of us should be using: category entry points. It comes from Professor Jenni Romaniuk and the Ehrenberg-Bass Institute, and the simplest way I can explain it is this. Category entry points are the cues that make a person think of buying in your category in the first place: what’s happening, where they are, who they’re with, what they need, how they feel. The important part is that they’re about the buyer, not the brand.
Some brands have taken the time to define these moments. A fizzy drink might say, “We show up when friends gather, we show up with food, we show up at the end of a long, hot day.” That’s the kind of thinking I mean, and it’s exactly how we should think about showing up in AI. Not “how do we rank for this keyword,” but “when someone describes this moment to an AI, are we a natural part of the answer?”
We don’t own the narrative
So how do we get into those conversations naturally? One of the favourite phrases in marketing is, “We want to own the narrative.” But even with all the contextual marketing we do today, we don’t own the narrative. I think it’s a bit prideful of us to believe we own what two people are saying to each other. At best, we join the narrative. We become part of the conversation. We’re not that important, and deep down we know it. Consumers are fickle. They like you today and go to someone else tomorrow.
AI seems to agree. Meltwater analysed more than 5 million citations across the major AI tools and found that 48% of everything they cite comes from news media. For ChatGPT, it’s 66%. When AI cites social platforms, Reddit makes up 28%, X 26% and YouTube 19%, the places where people discuss and explain things. Instagram is under 3%. TikTok is under 1%. Meltwater’s reading is that AI values conversation and consensus over aesthetics. Heineken’s global measurement lead put it better than I can: “It’s no longer a battle of keywords, but a battle of sources.”
AI seems to agree. Meltwater analysed more than 5 million citations across the major AI tools and found that 48% of everything they cite comes from news media. For ChatGPT, it’s 66%. When AI cites social platforms, Reddit makes up 28%, X 26% and YouTube 19%, the places where people discuss and explain things. Instagram is under 3%. TikTok is under 1%. Meltwater’s reading is that AI values conversation and consensus over aesthetics. Heineken’s global measurement lead put it better than I can: “It’s no longer a battle of keywords, but a battle of sources.”
What AI tools cite
Share of citations from news media: all major AI tools
48%
Share of citations from news media: ChatGPT
66%
Social citations: Reddit
28%
Social citations: X
26%
Social citations: YouTube
19%
Social citations: Instagram
<3%
Social citations: TikTok
<1%
Source: Meltwater, Own Your AI Visibility (2026) · GenAI Lens analysis of 5M+ citations
So AI isn’t going to repeat what we say about ourselves. It’s going to repeat what other people say about us. That’s the conversation we have to earn our way into.
Culture works the same way. You can’t own it, you can only be part of it. And when I look at what that means for marketing over the next few years, I see six shifts.
Culture works the same way. You can’t own it, you can only be part of it. And when I look at what that means for marketing over the next few years, I see six shifts.
Chapter 07
Six shifts I see coming
And it matters commercially.
3.6x
When an AI answer names a brand, that brand’s website visit rate goes from 1.8% to 6.8%, about 3.6 times higher, and that held across all nine industries they looked at.
1. Discovery: social is still winning
The consumer journey hasn’t collapsed. People still discover brands on social media, still watch influencers, still search for products, still ask “what’s good?” in the places they’ve always asked. That part hasn’t changed as dramatically as some people want to claim, and Gartner’s 25% search drop that never came is proof of it.
In Nigeria, 83% of people have bought something directly through social media, and 68% say they usually discover new brands or shops while browsing online. Social is still where people meet you.
But people don’t fully believe what they see there. In Gener8’s UK trust survey, social media and influencers came last of every source, trusted by only 29% of people. Billboards were at 32%, TV advertising at 42%, AI assistants at 47%, search engines at 75% and friends and family at 87%. So people find you on social, and then they go and check. More and more, they check with AI, which is now trusted more than every form of advertising on that list.
In Nigeria, 83% of people have bought something directly through social media, and 68% say they usually discover new brands or shops while browsing online. Social is still where people meet you.
But people don’t fully believe what they see there. In Gener8’s UK trust survey, social media and influencers came last of every source, trusted by only 29% of people. Billboards were at 32%, TV advertising at 42%, AI assistants at 47%, search engines at 75% and friends and family at 87%. So people find you on social, and then they go and check. More and more, they check with AI, which is now trusted more than every form of advertising on that list.
Who people trust
Social media and influencers
29%
Billboards
32%
TV advertising
42%
AI assistants
47%
Search engines
75%
Friends and family
87%
Source: Gener8 Survey · UK adults, September 2026, n = 1,020
2. Where AI has changed things: comparison and decision
What has changed is the comparison and decision stage. People go into AI and ask things like, “I want to buy a skincare product, what would you recommend based on what you know about me?” Or simply, “Choose one for me.” The decision to buy is increasingly made with AI in the room. Not only informed by it. Made with it.
The traditional funnel assumed a person was thinking through every stage, and we could influence them at each step. If the comparison and the decision are now handled by an AI pulling together everything it knows about the category, the products and the person asking, then the question becomes: how do you market to a funnel where part of the funnel isn’t human?
Gener8 summed it up in one line. The brand manager’s question used to be, “How do I make the consumer choose my brand?” Increasingly, it’s, “How do I make the AI choose my brand?” And it matters commercially. When an AI answer names a brand, that brand’s website visit rate goes from 1.8% to 6.8%, about 3.6 times higher, and that held across all nine industries they looked at.
The traditional funnel assumed a person was thinking through every stage, and we could influence them at each step. If the comparison and the decision are now handled by an AI pulling together everything it knows about the category, the products and the person asking, then the question becomes: how do you market to a funnel where part of the funnel isn’t human?
Gener8 summed it up in one line. The brand manager’s question used to be, “How do I make the consumer choose my brand?” Increasingly, it’s, “How do I make the AI choose my brand?” And it matters commercially. When an AI answer names a brand, that brand’s website visit rate goes from 1.8% to 6.8%, about 3.6 times higher, and that held across all nine industries they looked at.
Website visit rate
Baseline
1.8%
AI answer names the brand
6.8% · 3.6x
Person names the brand in their question
9.6% · 5.4x
Source: Gener8 GenAI and Clickstream data · US and UK, Aug 2025 to Jul 2026, 9 verticals
3.6x
When an AI answer names a brand, that brand’s website visit rate goes from 1.8% to 6.8%, about 3.6 times higher, and that held across all nine industries they looked at.
Brands need to start studying this behaviour seriously. Not only what people see, but what AI sees when it looks at your brand. Are you in the right conversations? Is the information about your product clear and accurate enough for AI to represent you well? When someone asks, “What should I buy?”, are you even in the answer?
3. Reach vs resonance: the metric problem
Reach and impressions have always been the numbers we chase. After every campaign, it goes in the report: 11.3M reach, 60M impressions, and someone says, “Everyone in the audience has seen the ad twice.” The client nods, and we move on.
Those numbers tell us how far something went. They don’t tell us what it did when it got there.
Attention is everywhere now, and it’s meaning that has become scarce. People can see your content and feel nothing at all. And feeling nothing is the new scroll past.
Bring AI into it and the problem gets bigger. When an AI decides which brands to recommend, it isn’t counting your impressions. It’s reading what people say about you, what’s written about you, and what you mean to people when there’s no campaign running. The brands people already carry in their heads win. Gener8 found that when someone names a brand in their question to AI, that brand’s visit rate jumps to 9.6%, 5.4 times the baseline. That’s resonance showing up in the numbers.
We need to move from reach to resonance. Not “how many people saw this?” but “what did it do to the people who saw it?” When you understand resonance, what made people stop, engage, share and come back, you understand your audience better, you understand what made the campaign work, and you give yourself the chance to repeat that success. Not just once. Consistently.
Those numbers tell us how far something went. They don’t tell us what it did when it got there.
Attention is everywhere now, and it’s meaning that has become scarce. People can see your content and feel nothing at all. And feeling nothing is the new scroll past.
Bring AI into it and the problem gets bigger. When an AI decides which brands to recommend, it isn’t counting your impressions. It’s reading what people say about you, what’s written about you, and what you mean to people when there’s no campaign running. The brands people already carry in their heads win. Gener8 found that when someone names a brand in their question to AI, that brand’s visit rate jumps to 9.6%, 5.4 times the baseline. That’s resonance showing up in the numbers.
We need to move from reach to resonance. Not “how many people saw this?” but “what did it do to the people who saw it?” When you understand resonance, what made people stop, engage, share and come back, you understand your audience better, you understand what made the campaign work, and you give yourself the chance to repeat that success. Not just once. Consistently.
4. Cultural intelligence: understanding beats awareness
I’ve worked in data insights and social listening for a long time, and cultural intelligence comes up a lot. Most of the time it means, “What’s trending and how do we jump on it?” That’s one way of seeing it, but it only covers the surface.
Cultural intelligence, to me, is understanding how a culture works from the inside. What are the codes? What are the references? What do people inside a particular culture or community respond to that people outside it simply can’t see?
This is where brands get it wrong. Something is trending, and two brands activate on it on the same day. One joins the conversation naturally. The other is awkward and out of place, and people either ignore it or make fun of it. And you’re left wondering what one brand did that the other didn’t.
Cultural intelligence, to me, is understanding how a culture works from the inside. What are the codes? What are the references? What do people inside a particular culture or community respond to that people outside it simply can’t see?
This is where brands get it wrong. Something is trending, and two brands activate on it on the same day. One joins the conversation naturally. The other is awkward and out of place, and people either ignore it or make fun of it. And you’re left wondering what one brand did that the other didn’t.
Sometimes it’s execution. Most times it’s understanding: whether you understand what that moment means to the people making it, or whether you’re just chasing the signal.
We see this all the time in the Nigerian market, brands pushed to jump on every trend and activate on everything, fast. Cultural intelligence also means knowing when to say, “This doesn’t fit us,” or, “This is too deep for us to make light of.” Very few people talk about that part.
In an AI world, this matters even more. AI repeats what people say about you. The brand that got dragged for a badly judged post doesn’t just lose a news cycle. That reaction becomes part of the record AI reads, the Reddit threads and X posts it cites most often. You can’t fake proximity, and you can’t prompt your way out of a bad reputation.
Duolingo gets mentioned a lot as a brand that does this well. The better question is what work happened behind the scenes to make it look so effortless. It’s never just the post. It’s the understanding that came before it.
We see this all the time in the Nigerian market, brands pushed to jump on every trend and activate on everything, fast. Cultural intelligence also means knowing when to say, “This doesn’t fit us,” or, “This is too deep for us to make light of.” Very few people talk about that part.
In an AI world, this matters even more. AI repeats what people say about you. The brand that got dragged for a badly judged post doesn’t just lose a news cycle. That reaction becomes part of the record AI reads, the Reddit threads and X posts it cites most often. You can’t fake proximity, and you can’t prompt your way out of a bad reputation.
Duolingo gets mentioned a lot as a brand that does this well. The better question is what work happened behind the scenes to make it look so effortless. It’s never just the post. It’s the understanding that came before it.
5. Community over culture: the asset AI can’t copy
Culture is a language we can’t escape in marketing. Read any brand strategy document and you’ll see it: “We want to inject the culture,” “We want to own the culture,” “We want to drive the culture,” “We want to create the culture.” After a while, those words lose their meaning.
The word that means more to me is community, especially now. As AI makes it easy for any brand to produce good-looking, well-written campaigns, good content stops being what sets you apart. A look can be copied. A real community can’t.
I mentioned Gen Z earlier. Audiences are not one fixed group. The deeper work is understanding what people like, who they are, and which communities they belong to. That’s where the insight lives.
Over the next few years, I think marketing will be done more and more on a community basis. Where do these people go? What are they watching on YouTube? Which podcasts are they in? Which creators are they in conversation with? Creator marketing will keep growing, that argument has been made. What will matter most is not which creator has the biggest following, but which creator has built something real with the people around them.
The data points the same way. The social platforms AI cites most, Reddit, X and YouTube, are the ones where people talk to each other. And when people were asked who they trust, friends and family came first at 87%. Community is what makes the impact last.
The word that means more to me is community, especially now. As AI makes it easy for any brand to produce good-looking, well-written campaigns, good content stops being what sets you apart. A look can be copied. A real community can’t.
I mentioned Gen Z earlier. Audiences are not one fixed group. The deeper work is understanding what people like, who they are, and which communities they belong to. That’s where the insight lives.
Over the next few years, I think marketing will be done more and more on a community basis. Where do these people go? What are they watching on YouTube? Which podcasts are they in? Which creators are they in conversation with? Creator marketing will keep growing, that argument has been made. What will matter most is not which creator has the biggest following, but which creator has built something real with the people around them.
The data points the same way. The social platforms AI cites most, Reddit, X and YouTube, are the ones where people talk to each other. And when people were asked who they trust, friends and family came first at 87%. Community is what makes the impact last.
6. Earned first: bravery before budget
I heard an idea recently that I’ve probably met before in other forms, but this time it settled for me: the earned-first approach.
The idea is simple, but it takes a kind of bravery most brands aren’t ready for. Instead of designing a campaign and then hoping it gets people talking, you build the campaign with earned media as the goal from day one. Before anything else, you ask: is this idea strong enough that people will want to share it, talk about it and engage with it without being paid to?
The example shared was Pedigree’s “Caramelo” campaign, made by AlmapBBDO for Mars in Brazil. The vira-lata caramelo, the caramel-coloured mutt, is the dog almost every Brazilian has grown up around. It’s a national symbol. But when people want a dog, they want a purebred, and the caramelos stay in shelters. Pedigree could have run a sweet adoption ad and stopped there. It went all the way in instead. It commissioned a genetic study of 350 caramelos to show they share physical, genetic and health traits, the way a breed does. It started a Caramelo Kennel Club. It held the first dog show for mixed-breed dogs, with an adoption fair attached. It even put a caramelo on its packaging. Pedigree gave people something to talk about, and they talked. The campaign won AlmapBBDO its first Titanium Lion at Cannes in 2025, and Pedigree says adoptions through its programme went up 218%. It’s worth watching, because it shows what happens when a brand commits fully to an idea instead of hedging.
The idea is simple, but it takes a kind of bravery most brands aren’t ready for. Instead of designing a campaign and then hoping it gets people talking, you build the campaign with earned media as the goal from day one. Before anything else, you ask: is this idea strong enough that people will want to share it, talk about it and engage with it without being paid to?
The example shared was Pedigree’s “Caramelo” campaign, made by AlmapBBDO for Mars in Brazil. The vira-lata caramelo, the caramel-coloured mutt, is the dog almost every Brazilian has grown up around. It’s a national symbol. But when people want a dog, they want a purebred, and the caramelos stay in shelters. Pedigree could have run a sweet adoption ad and stopped there. It went all the way in instead. It commissioned a genetic study of 350 caramelos to show they share physical, genetic and health traits, the way a breed does. It started a Caramelo Kennel Club. It held the first dog show for mixed-breed dogs, with an adoption fair attached. It even put a caramelo on its packaging. Pedigree gave people something to talk about, and they talked. The campaign won AlmapBBDO its first Titanium Lion at Cannes in 2025, and Pedigree says adoptions through its programme went up 218%. It’s worth watching, because it shows what happens when a brand commits fully to an idea instead of hedging.
Pedigree “Caramelo” by AlmapBBDO for Mars, Brazil
Every time this comes up in a meeting, I ask myself the same question. We all say we want earned media. It’s in every brief. But are we designing campaigns brave enough to earn it? The bravery isn’t in the execution. It’s in the thinking that comes before it, in being willing to build something that might not work the safe, predictable way, because when it does work, your audience does the work for you.
In an AI world, earned media isn’t a bonus. It’s the strategy. Almost half of what AI cites comes from news media, and AI prefers what’s recent: in Meltwater’s analysis, 58% of the dated sources AI cited were less than a year old. As Allie Konchar, co-founder of Omniscient, puts it in the same guide, earned media creates exactly the kind of independent, high-authority sources AI relies on when it summarises and recommends brands. Paid media gets you seen. Earned media gets you into the answer.
That’s harder to pitch internally than a media plan. It’s also much harder to lose.
In an AI world, earned media isn’t a bonus. It’s the strategy. Almost half of what AI cites comes from news media, and AI prefers what’s recent: in Meltwater’s analysis, 58% of the dated sources AI cited were less than a year old. As Allie Konchar, co-founder of Omniscient, puts it in the same guide, earned media creates exactly the kind of independent, high-authority sources AI relies on when it summarises and recommends brands. Paid media gets you seen. Earned media gets you into the answer.
That’s harder to pitch internally than a media plan. It’s also much harder to lose.
Chapter 08
So what should marketing do?
I don't think marketing suffers from a lack of data. If anything, it suffers from an abundance of it.
Lead with your heart, but make your decisions with the numbers.
Truthfully, I don’t know. I never set out to write a knowledge piece with five steps you should follow. But Meltwater’s Own Your AI Visibility guide helped me start thinking about how a brand can take some control of how it shows up in AI, and a few things have stayed with me.
You have to start by accepting that AI has changed how the world finds out about your brand, how people engage with it, and how they decide. Once we accept that, we can think properly about how to respond.
You have to start by accepting that AI has changed how the world finds out about your brand, how people engage with it, and how they decide. Once we accept that, we can think properly about how to respond.
Audit everything
The first step, from the research I found, is to audit. Audit your current visibility, audit your positioning, audit everything you have. And not as a box-ticking exercise. Make it a real, deep one, because otherwise you won’t understand your audience. What are they doing? What are they thinking about? How are they responding to everything happening around them? And how did they find you when they did their research?
A lot of brands I know don’t even know their search data: whether they come up in search, which topics they come up for, or where. With some consumer brands, you assume people already love them, the loyalist kind of conversation. But when we think about human behaviour, it isn’t always about preference. It’s also about life choices. Someone might be asking Google or ChatGPT, “What beer can I drink that won’t make me burp, won’t make me throw up and won’t give me a big stomach?” That’s how people actually search, and that’s how you should think about it. Is your brand in that answer?
It’s also worth knowing where AI gets its information. Meltwater found that up to 45% of AI citations come from places many brands neglect: their own blogs and content hubs, product FAQs and help pages, review sites and forums, and niche industry publications. AI also favours content it can easily pull apart, like clear answers, comparisons and lists, over long blocks of marketing language. Your audit should look at all of that, not only your ads and your social pages.
A lot of brands I know don’t even know their search data: whether they come up in search, which topics they come up for, or where. With some consumer brands, you assume people already love them, the loyalist kind of conversation. But when we think about human behaviour, it isn’t always about preference. It’s also about life choices. Someone might be asking Google or ChatGPT, “What beer can I drink that won’t make me burp, won’t make me throw up and won’t give me a big stomach?” That’s how people actually search, and that’s how you should think about it. Is your brand in that answer?
It’s also worth knowing where AI gets its information. Meltwater found that up to 45% of AI citations come from places many brands neglect: their own blogs and content hubs, product FAQs and help pages, review sites and forums, and niche industry publications. AI also favours content it can easily pull apart, like clear answers, comparisons and lists, over long blocks of marketing language. Your audit should look at all of that, not only your ads and your social pages.
Make the changes strategically
The next thing is to make changes strategically, as they become important. I don’t think we’ll ever get to a point where everyone is 100% right or fully on top of what’s happening. AI models change constantly. We’ll have to learn and adjust as we go. That takes strategic capacity: what do we want to achieve in the next two to four months? What about the next six to twelve?
Stop collecting data for data’s sake
The third one matters a lot to me, because it’s the world I work in. We have to stop collecting data for data’s sake. There’s a reason these numbers are shared. There’s a reason this research gets put together. It’s meant to change something in us, to help us think differently.
I don’t think marketing suffers from a lack of data. If anything, it suffers from an abundance of it. There’s so much data available, and marketers use only a small part of it. What would help is understanding the numbers. What is the data saying? What are people saying? Lead with your heart, but make your decisions with the numbers. There’s a heart side and a head side to this work, and data serves the head.
I don’t think marketing suffers from a lack of data. If anything, it suffers from an abundance of it. There’s so much data available, and marketers use only a small part of it. What would help is understanding the numbers. What is the data saying? What are people saying? Lead with your heart, but make your decisions with the numbers. There’s a heart side and a head side to this work, and data serves the head.
Be curious again
Finally, be curious. I feel like the fun part of marketing is going away, the part where we spent time finding things out.
I remember my early years in marketing, and this is a random example. Whenever we were planning an event, one of the first things we did was go to the venue. We called it a recce. We took tape rolls, measured everything and took pictures. We were curious about the experience people would have. We looked at the area around it. Is there good parking? What could we do on the street outside?
I remember my early years in marketing, and this is a random example. Whenever we were planning an event, one of the first things we did was go to the venue. We called it a recce. We took tape rolls, measured everything and took pictures. We were curious about the experience people would have. We looked at the area around it. Is there good parking? What could we do on the street outside?
A venue recce: walking the room with the floor plan
That same curiosity ran through the rest of our work. When we needed research, we interviewed people, asked questions and ran focus groups. I haven’t done that in a long time, and I think it’s because curiosity is leaving marketing. Everyone is a professional marketer now, even the CFO, and we’re losing the joy of asking, “What if we tried this?”
I’d add one more thing. We marketers need to let go of the idea that we only do things because we already know them, or because we’re certain how they’ll turn out. There’s so much pride in our work that we forget half of it is testing. Will this work? How will the audience react? Think about the biggest campaigns you can think of, the ones that won every award. They weren’t built from certainty. They were built from, “We understand this audience deeply, and we believe this will connect.” Then someone was brave enough to put the work out. Bravery needs to come back to marketing, and that’s how we win in this AI era. The brands that win will be the ones that think differently, use their data well, stay curious enough to want to do great work, and are brave enough to ship it.
I’d add one more thing. We marketers need to let go of the idea that we only do things because we already know them, or because we’re certain how they’ll turn out. There’s so much pride in our work that we forget half of it is testing. Will this work? How will the audience react? Think about the biggest campaigns you can think of, the ones that won every award. They weren’t built from certainty. They were built from, “We understand this audience deeply, and we believe this will connect.” Then someone was brave enough to put the work out. Bravery needs to come back to marketing, and that’s how we win in this AI era. The brands that win will be the ones that think differently, use their data well, stay curious enough to want to do great work, and are brave enough to ship it.
Chapter 09
To close
This is my first-ever thought piece on marketing, after almost ten years of working in it.
I’m excited about the future of marketing.
This is my first-ever thought piece on marketing, after almost ten years of working in it. I usually write about other things, and this is probably the longest piece I’ve ever written.
I’m excited about the future of marketing. I think we’re at a point where marketing can do more good in the world than we realise, and I’m very pumped to be in this space.
I hope you enjoyed my first piece. I don’t know where I’ll share the next one, but thank you for reading.
I’m excited about the future of marketing. I think we’re at a point where marketing can do more good in the world than we realise, and I’m very pumped to be in this space.
I hope you enjoyed my first piece. I don’t know where I’ll share the next one, but thank you for reading.
How this was written
- The ideas started at the Meltwater Summit, which I attended.
- Every thought in this piece is mine. I spoke them out loud and captured them with Wispr Flow.
- AI helped me edit and shape them.
- The images and illustrations were designed with AI.
- The final edit was mine.
Sources
01
Gener8, How Our AI Use Changed in a Year (September 2026) and What 100 Million Real AI Conversations Reveal About Your Customers (October 2026). US and UK consented panel data, August 2025 to August 2026.
02
Meltwater, Own Your AI Visibility: Shape What AI Says About Your Brand (2026).
03
Google and Ipsos, Our Life with AI, via Techpoint Africa.
04
Visa, Stay Secure 2026, via TheCable.
05
Orange and KPMG, Nigeria Smartphone Study 2025, via Brand Spur.
06
Gartner’s 2024 search prediction and what happened, via Future Factors.
07
Levenson et al., PLOS ONE (2015), the pigeon study, via Phys.org.
08
Ehrenberg-Bass Institute on category entry points, via Marketing Week.
09
Union Bank’s rebrand and “Don’t be an Uncle Thomas”: Guardian Nigeria, Ads of the World, YNaija.
10
Pedigree “Caramelo” by AlmapBBDO: LBB Online, Ads of the World, Propmark (adoption figure).
If you want to think this through with us
I wrote this in my own name, but a lot of how I see it comes from the work we do at ImpressionsON. We spend our days on social listening, consumer research and analytics, trying to understand why people click, buy and talk the way they do. AI is now part of that picture.
If you're trying to work out how AI sees your brand, what your audience is asking AI about your category, or what to change first, we'd be glad to talk. It doesn't have to start as a big project. Often it starts with an audit and a conversation.
If you're trying to work out how AI sees your brand, what your audience is asking AI about your category, or what to change first, we'd be glad to talk. It doesn't have to start as a big project. Often it starts with an audit and a conversation.
Tell us what you need