I ran a brand I know well through an AI visibility check last week. It does fine when someone asks "what's the best tool for X." Cited. Described accurately. Not bottom of the list.
Then I ran a different query. Not "what's the best tool." Just the question one of its actual buyers would type before they knew the category existed. Something like "why is my onboarding taking three weeks when competitors do it in three days."
Nothing. Not a mention. Not even a wrong description. The brand didn't exist for that question.
That's the gap I want to write about.
What the research actually shows
There's research floating around on AI search visibility that quantifies this gap. The headline finding the industry has been chewing on is around the 71% mark.
Reading it the way I read it, roughly 71% of brands that show up for category-led queries vanish when buyers ask the same thing as a problem. Same brand, same product, same buyer. Different phrasing of the question. And most of them disappear.
That should be a five-alarm fire for anyone running content strategy.
It isn't, mostly because content strategy has spent twenty years optimising for the wrong shape of question.
How buyers actually search before they know what to buy
Think about how you research something you don't fully understand yet.
You don't type "best CRM for mid-market B2B SaaS." Not at first. You type "why is my sales team spending so much time on admin." You type "how do I track which leads my reps are ignoring." You type "what's the thing called when leads don't come back to you after a demo."
You're not searching for a category. You're describing a problem. The category is the answer you don't have yet.
Old search rewarded the second stage of that research. The category stage. "Best CRM for X." Pages and pages of comparison content optimised for those queries. SEO trained a generation of marketers to write for the buyer who already knew what they were buying.
AI search rewards the earlier stage. Problem-led queries get conversational answers. The model doesn't bounce you to ten blue links. It tells you what's probably going on and which kinds of products solve it. If your brand isn't in the answer to "why is my sales team spending so much time on admin," you're not in the running by the time the buyer gets to "best CRM."
You lost before the comparison started.
Why category-led content is the default trap
Most content strategies I look at have a category-shaped backbone.
There's a pillar page on "what is a CRM." There are comparison pieces. There's a buyer's guide. There's a "best of 2026" listicle. There's an integrations page. The whole thing assumes the buyer arrives knowing the word CRM and shopping for one.
I wrote pages like that for years. I'd rebuild a client's site, structure it around their service categories, and call it strategy. It worked when Google rewarded keyword density. It mostly still works for branded queries.
But it leaves the buyer's actual research moment uncovered. The moment when they're typing the problem in their own words, before they've earned the vocabulary to describe what they need.
If your only content is category content, you've optimised for the buyer who already knows you exist. That's not who you need to reach.
What problem-led content looks like in practice
A few rules I've started using when I rewrite a client's content plan.
Lead with the problem in the buyer's words, not yours. If a buyer would call it "leads going cold after the demo," call it that. Don't translate it into "post-demo conversion drop-off." The model picks up what humans actually say, not what marketers think humans should say.
Name the symptom before the solution. The first paragraph of the page should sound like a diagnosis, not a pitch. "If you're seeing X, here's probably what's going on." That mirrors how AI answers a problem query. It's also how the buyer expects to be met.
Stop hiding the category until late in the page. Most SEO playbooks tell you to put the keyword in the first sentence. For problem-led pages, the keyword for the category often shouldn't appear until two-thirds of the way down. The page is doing diagnostic work first. The category name is a label for the conclusion, not the headline.
Write the questions the way a buyer would say them out loud. FAQ sections are AI gold when the questions are honest. "Why isn't my landing page converting" beats "How to improve landing page conversion rate." One sounds like a person. The other sounds like a content brief.
The studio thesis underneath this
Level44's strapline is "for brands designed to be seen." I keep coming back to it because the implication shifts depending on which stage of buyer research you're talking about.
Being seen at the category stage is fine. It's table stakes. By the time a buyer is comparing options in your category, the shortlist is mostly already drawn. You're fighting for a spot in a small set.
Being seen at the problem stage is the whole game. That's where the buyer is forming the shortlist. That's where they're learning the vocabulary of the category. That's where they're deciding whether the problem they have even has a category that solves it.
A brand that's named at the problem stage is the brand that frames the buyer's understanding of what they need. By the time they get to "best X for Y," they're already half-sold on the brand that named the problem accurately for them in week one of their research.
That's worth more than every comparison page on your site combined.
Where this leaves the next twelve months
I think most brands have a small window. The 71% problem isn't going to last forever. The brands that figure it out in the next year are going to own the problem-stage answer for their category.
Practically, that means a content audit that's brutal about category vs problem. List every piece of content you have. Mark the ones that lead with a category and the ones that lead with a problem in the buyer's words. If the second list is shorter than the first, that's the work.
Then think about which problems you actually want to be the named answer for. Not "the brand that comes up when someone searches for our category." That's already a contested space. The named answer for the problem that comes before the category gets searched. That's the position worth fighting for.
It's not a content tactic. It's a brand positioning question dressed up as a content one.
The brands that get this right won't just be cited in AI answers. They'll be the brand the buyer was already thinking of by the time they got there.