Back in January, we said trying to predict 2026 felt a bit like reading a map while the ground shifted underneath you. Now we are the far side of the summer, and it turns out the ground has been busier than usual.
Some of what the big reports pointed to has arrived faster than anyone expected. Some of it has settled into the background so quietly you would barely notice. And a few things have taken a turn that almost nobody had on their radar in the new year.
So, coffee in hand again, here is what we are actually seeing play out with our clients now that we are over the halfway mark, and where we think the smart money is heading next.
AI stopped being the story and became the gatekeeper
We said at the start of the year that AI had moved from shiny new thing to plumbing. That has held. What we did not fully call was how quickly the mood would sour on lazy AI content, and how fast AI would start making decisions on the audience’s behalf.
The backlash is real and it is measurable. Gartner found nearly half of consumers now think AI has made content quality worse, and appetite for AI-generated creator content has dropped off a cliff, from around 60% in 2023 to roughly a quarter today, according to reporting pulled together by Digiday. “AI slop” has become a genuine term of art, with a fifth of the videos served to new YouTube users reportedly falling into that bucket. Audiences can smell it, and they are voting with their thumbs.
AI isn’t going anywhere, but the way we use it will change. You get what you pay for: bold, striking work still needs skilled people making the calls.
The bottom end of the industry will keep leaning on it for cheap, quick jobs but big brands simply can’t get away with taking that shortcut. Skilled professionals have to decide how much of the process they hand over.
At Flourish, AI speeds things up. It doesn’t make the decisions. The ideas and the strategy stay ours, and that’s where the value has always sat.
At the same time, AI has quietly become a buyer. Kantar’s mid-year work points to a shift “from attention to intention”, with roughly a quarter of AI users already leaning on an AI shopping assistant and the majority actively seeking out AI recommendations. The practical upshot is a new discipline sitting next to SEO, generative engine optimisation (GEO) – sometimes called LLM SEO -, where the job is making sure your brand is legible to the machine doing the recommending, not just the human doing the reading.
What this means in practice
Two jobs now, not one. Keep the human judgement in your content so it does not read as slop, and make sure your brand is structured, clear and consistent enough that an AI can find it, understand it and recommend it.
AI Overviews and zero-click did not creep in. They took over.
If we underplayed anything in January, it was the pace of this one. We said AI search and zero-click were not a passing phase. What has actually happened is that they have become the default.
Recent SparkToro analysis of Similarweb data put the share of Google searches ending without a click at around 68% in the first few months of 2026, up from roughly 60% two years ago. AI Overviews now appear on a large slice of searches and cut click-through on those results dramatically. Google’s AI Mode has passed a billion monthly users. Closer to home, Yext found more than a third of UK consumers had used AI for a local search in the past month, and three quarters were using it more than they were a year ago.
The encouraging part for anyone who has invested in being genuinely useful: people still verify. That same UK research shows the top actions after an AI recommendation are checking Google and visiting the brand’s own website. The click has not died. It has just moved to the end of the journey rather than the middle. So the job changes shape. You are no longer only competing to be clicked; you are making sure that when the AI describes you, and when the person double-checks, everything they find quietly backs it up.
What this means in practice
Treat the AI answer as your new shop window. Structure your site, FAQs and service pages around the real questions people ask, keep your listings and reviews saying the same thing everywhere, and make it effortless to act the moment someone does land, so the visit that still happens actually converts.
“Less perfect, more human” hardened into strategy
This was our steadiest prediction, and it has aged well. The move towards real people and honest content has not just continued, it has become a proper line in the budget.
In the UK specifically, the influencer marketing numbers are striking. Kolsquare’s figures, reported by Marketing Week, show around four in five UK brands increasing their influencer budgets this year, and six in ten shifting towards longer-term partnerships rather than one-off posts, exactly the direction we flagged. The wider creator economy is maturing fast; it is now a multi-billion-pound industry in Britain with tens of thousands of jobs behind it. The style has shifted too, towards what some are calling “mid-fi” content: less gloss, more behind-the-scenes, the odd unmade bed left deliberately in shot.
The one note of caution is accountability. Kantar’s mid-year read is blunt that creators now have to earn their place at the effectiveness table, pointing out that only around a quarter of creator content is strongly tied to the brand paying for it. It is telling, too, that some platforms are now testing hiding follower counts altogether. The measure of a good partnership is quietly shifting from how big an audience a creator has to whether they have built a genuinely engaged community around it. Warmth is not a substitute for relevance, and reach is not a substitute for engagement.
We saw this play out ourselves this year with Camp Connect USA, when we launched paid US summer camp jobs to a UK audience for the first time. We took a micro-influencer marketing approach, choosing real camp counsellors and ambassadors and let them lead, rather than chasing the biggest following. They worked in the trends and formats their audience already watched. We briefed them to keep it simple and honest: show a normal day-in-the-life, answer the questions their friends actually ask and skip the scripts. It worked. Giving people a genuine, aspirational but realistic sense of what to expect made the call to apply land far better than a polished pitch. We lead with real voices first because people trust people over adverts.
What this means in practice
Pick fewer creators and stay with them longer. Judge them on the community they have built and how it engages, not the size of their following, and be clear about what your brand stands for so the content is memorable for the right reasons.
Communities got smaller, and moved somewhere quieter
We talked in January about micro-communities and real-world moments. Six months on, that has accelerated, and a lot of the action has moved off the open feed entirely.
Audiences, younger ones especially, are spending more time in closed and semi-closed spaces like Discord, Substack and group chats, where reach is smaller but trust is far higher. Kantar found nearly four in ten people now trust a recommendation from a micro-community roughly as much as one from someone they know personally. With ongoing uncertainty around the big platforms, “own the room you are in” has started to look a lot wiser than “be everywhere”.
The brands getting the most from this are not stopping at the screen, either. They are bringing those communities into the real world, hosting meet-ups, workshops and events that give members something genuinely worth turning up for. An engaged online community is the start; a reason to gather in person, and real value for being part of it, is what turns a following into loyalty.
What this means in practice
Find the two or three places your audience already gathers and show up there properly, as a member rather than a billboard. Then look for ways to take it offline, with IRL events and experiences that reward people for being part of it. Depth is beating breadth.
A new economic mood: comfort, treats and the vanishing middle
Here is something that was barely on the agenda in January and is now hard to ignore. WARC’s recent work describes a “vanishing middle”, with years of economic pressure reshaping not just how people spend but the life milestones they aim for. Kantar frames the flip side as “Treatonomics”: people denying themselves the big things and rewarding themselves with small, affordable pleasures instead, with a surprising share willing to stretch a little to do it.
This is where the nostalgia marketing we wrote about earlier in the year meets the current climate. The appetite for comfort, familiarity and small joys has not faded; it has found an economic rationale. You can see it in who is winning. Lego has built a genuinely grown-up business on nostalgia, with elaborate adult sets aimed squarely at the childhood memories of thirty- and forty-somethings. Motorola has turned the return of the Razr flip phone into one of its fastest-growing lines, and Pizza Hut’s retro dine-in reboots have become some of its best-performing sites. In beauty, the old “lipstick index” is playing out in a modern form, with an affordable lip oil, a scent or a small treat standing in for the big-ticket purchase people are holding off on. The common thread is not the decade being referenced; it is that each one offers a real, well-priced hit of comfort. Brands using nostalgia as empty set-dressing, by contrast, still are not landing.
What this means in practice
Read the room. People are careful with money but still want to feel good. A small, sincere, well-judged treat, or a genuine moment of comfort, will do more for you right now than a grand gesture.
What we are watching next
A few things are on our radar for the back end of the year and into 2027. First, agentic commerce moving from novelty to habit; as more people let AI assistants shortlist and even buy, being recommendable becomes a core marketing skill, not a technical afterthought. The same shift rewards a related move, the industry aggregator: AI makes it far easier to pull a fragmented market into one genuinely useful place, think every local theatre’s listings in a single view instead of six separate websites, and a brand that becomes that one-stop resource for its category has an easy win on its hands. Second, a measurement reckoning, as leaders push creators, communities and AI-assisted content to prove they actually build the brand. And third, a continued flight to the human and the owned, from email lists and communities to events and real relationships, as rented reach on the open platforms gets noisier and less certain.
And keep one eye on AI itself sliding down the funnel. So far the AI conversation has mostly been about discovery, being found, understood and recommended. That is starting to change. ChatGPT ads moved into self-serve beta in May and reached the UK over the summer, complete with conversion bidding and a tracking pixel, and early Criteo data suggests traffic referred from AI tools is converting at around one and a half times the rate of other channels. The shift worth watching is LLMs moving from the top of the funnel to the bottom, from where people get their ideas to where they actually buy. If that holds, the same discipline that gets you recommended will need to get you bought.
Final thoughts
The thread running through January’s piece was intention, and if anything the case for it has got stronger. The tools have got faster, the feeds have got noisier and the machines have started making choices for people. None of that rewards doing more for its own sake.
The brands doing well halfway through 2026 are the ones being deliberate: clear about where they show up, honest about how they sound, useful enough to be recommended, and human enough to be trusted. That was true in January. It is even more true now.
