When the internet first took off, people did not know how to navigate it. Early solutions looked like what we already understood offline: directories, curated lists, the phone book on a screen. Then Google changed the front door. Instead of browsing, you could ask. And once that happened, everything else reorganized around it.
AI chat experiences are another front door change.
They are not just improving search. They are changing how people discover information, how they make decisions, and how brands influence those decisions. Queries are getting longer. More of them are informational. Fewer clicks are happening in the moment. Yet influence is still very real.
That creates a problem for marketers.
The old playbook was built around links, clicks, and last-touch conversion. The new reality is built around answers, discovery, and decisions that often happen without a click at all.
In this article, we unpack how to operate inside that shift, or watch it live below.
We will look at how search behavior is changing, how brands get discovered in AI-driven experiences, where ads and commerce fit without breaking trust, and how to think about measurement when clicks are no longer the main signal.
Search behavior didn’t evolve. It flipped.
One of the most important things to internalize is that AI didn’t just make search “better.” It changed how people search.
If you think about how you use ChatGPT, Gemini, Perplexity — you’re not typing “running shoes” anymore. You’re typing full sentences. You’re giving context. You’re asking questions the way you’d ask a smart friend.
That’s showing up in the data too: the fastest-growing bucket of queries is six+ word searches. People are entering full prompts, not keywords.
And when queries get longer, they also get more informational.
That matters because for the last twenty years, search marketers (paid or organic) had one shared belief: search is bottom-funnel. Even when we argued about everything else in the funnel, we agreed search lived close to conversion.
AI has flipped that.
A lot of what’s happening in AI search is upper-funnel discovery — informational, exploratory, “help me decide” behavior. And here’s the part that makes marketers uncomfortable:
A lot of it is zero-click.
People are getting answers and making decisions without ever visiting a website in the moment. That doesn’t mean there’s no marketing impact. It means the impact is showing up differently — often later — and we need to learn how to measure and influence it.
So how does a brand “rank” on an LLM?
This is the question clients are asking more than any other right now.
And I’ll start with the least exciting answer because it’s also the most important: traditional SEO still matters. A lot.
Not only because the best practices still apply — clean technical foundations, structured content, clear information architecture — but because LLMs need reliable, well-organized source material. Garbage in, garbage out is still the rule.
Put the most important information up top. Be clear. Make the takeaways easy to extract. Don’t bury the point in paragraph eight. (All of us have done this. It’s fine.)
But there’s another layer now too: the world outside your website.
LLMs are influenced by multimodal and third-party sources — and each model weights them differently. That can include:
- third-party sites and listicles
- news coverage
- YouTube content
- community discussion (yes, often Reddit — though it’s evolved in importance over time)
And here’s the thing that’s actually useful: even though LLM ranking systems are also black boxes, they do show their work in one key way: citations.
So the first step to improving visibility isn’t guessing. It’s investigating.
This is where I get to sound more exciting than my job probably is: you need to become a digital detective.
If you’re seeing a competitor consistently cited for a topic and that citation is coming from Reddit, that’s a different action path than if it’s coming from a listicle, which is different again than if it’s coming from the competitor’s own site.
Same outcome (they’re visible), totally different levers (how they got there).
That’s the practical starting point: reverse engineer the sources that are being rewarded, then build a plan to earn presence in those source types.
This is also where vertical context starts to matter. The sources and signals that drive AI visibility for a retailer look very different from those that matter for a financial services brand. That’s why we’ve built vertical-specific AI visibility guides because the “how” of showing up in LLMs is increasingly category-dependent, even if the underlying principles are shared.
Read your industry specific guide: Retail + eCommerce | Finance | Education | Travel + Hospitality | B2B
Content refresh is the low-hanging fruit everyone ignores
Marketers love new ideas. We love the “wow” factor. We want to reinvent the wheel because it feels like progress.
Sometimes the highest impact move is not the sexiest one: refresh what you already have.
Speed of refresh matters more in an AI discovery world, because this ecosystem moves fast and gets reweighted constantly. A good rule of thumb we’ve aligned on: refresh relevant, time-sensitive content every three months.
If your content includes stats, market context, rates, comparisons, product details, anything that can go stale, quarterly refreshes are a strong baseline.
What’s telling is where AI content platforms are focusing right now. Tools like AirOps are leading not with flashy features, but with a very practical promise: automate content refresh so your best work does not quietly decay. That shift in positioning says a lot about where the real leverage is.
Refreshing content is the fastest way to win ground, especially compared to trying to rebuild a broken SEO foundation from scratch.
Ads are coming to AI experiences. The only question is what form they take.
Let’s just be blunt: yes, ads are coming.
The internet is powered by ads. Subscriptions will play a role, but they won’t carry the full weight, especially not at the scale these products are chasing. The interesting question is not whether ads appear. It’s how they appear without breaking the experience that made LLMs popular in the first place.
Because the moment ads feel intrusive, biased, or like they’re blocking the answer, users will revolt.
We have already seen this play out. Perplexity launched sponsored suggested follow-up questions and CPM-based ad models, and then pulled them back quickly after user backlash. Even when something is not technically an ad, if it feels sponsored or manipulative, users react immediately.
That is the tightrope these platforms are walking. Monetize, but do not ruin trust.
This is why Google won the search ad wars in the first place. Quality Score wasn’t just an auction mechanic. It was an experience protection system. It helped ensure ads weren’t simply “who paid most,” but “who is most relevant,” so users didn’t hate the product.
LLMs will need an equivalent, or something even stricter. User expectations are higher. People came to these experiences to remove friction, not to be sold to.
For brands, that means the work does not start when ads finally launch. It starts now. We have laid out what that preparation looks like, from tightening inputs to rethinking relevance and measurement, in a separate piece on the future of LLM Advertising.
Measurement when clicks disappear: what do we track?
If you’re a performance marketer, “zero-click” is where the anxiety kicks in.
If the user doesn’t click, CTR doesn’t matter the same way and the conversion path doesn’t look clean. So what do you measure?
We’re seeing three metrics categories emerge as most useful:
- AI Share of Voice: How often do you show up on the prompts that matter?
This parallels SEO thinking, but there’s a twist: prompts are longer and more varied. There aren’t always obvious “tentpole” keywords the way there were in classic search.
- Volume of mentions
A lot of third-party tools over-index on relative scoring (“how you perform vs competitors”), and they often bury the raw volume.
But volume matters because ranking 85% of the time on ten prompts is not the same business impact as ranking 15% of the time across a million prompts.
- Correlated impact signals
Referral traffic matters, but it will be limited for obvious reasons.
So you also look for correlated signals: when mentions spike, do you see:
– an uptick in direct traffic
– an uptick in branded search
– an uptick in brand-driven conversions downstream
It’s not perfectly scientific out of the gate, but it’s directionally powerful — and it mirrors how we’ve always dealt with other upper-funnel and low-click channels like CTV, podcasts, and programmatic awareness.
One of the hardest parts right now is defining what to track.
If you’re a brand like a luxury apartment platform, your old tentpole might be “luxury apartment.” But no one goes into ChatGPT and types “luxury apartment.” They type: “best apartment in northwest DC with a Peloton” or “nicest place in southwest DC with a clean modern aesthetic.”
That means the work is no longer about tracking a handful of keywords. It is about defining the right universe of prompts, understanding intent within that universe, and measuring visibility in a way that connects back to real business outcomes rather than vanity metrics.That’s where we’re spending time with our clients right now.
And candidly? A lot of third-party “AI visibility scores” are lazy. They’re mostly competitor-relative and backward-looking.
What brands actually care about is: what is driving business for me? And where is demand going next?
We have gone deeper on this in a separate piece focused on the new metrics that matter in AI and zero-click environments, including how to think about visibility, influence, and downstream impact when clicks are no longer the main signal.
And this part is worth internalizing. We’re going to get new “AI insights” inside Google Analytics and Google Ads soon. Everyone will get excited. We’ll open them. They won’t be what we need. Then competitors will release different data. And we’ll play the most fun game in advertising: telling platforms what they need to build because someone else is doing it better.
The resource question: discovery vs conversion isn’t a clean split anymore
A lot of teams still feel pressure to justify search spend the old way: “every dollar in, X dollars out.”
Google itself trained the market on that sales pitch. And in some ways, they painted themselves into that corner.
But now the opportunity is bigger than bottom-funnel capture. There’s a massive chance to own the mid-funnel and we can’t count on clients to magically understand that value on their own. We have to show it. And we definitely can’t count on the platforms to tell that story for us.
The real shift is recalibrating what “commercial intent” means.
A prompt like “Can I get around Disney World without renting a car?” doesn’t sound commercial. But if you’re a hotel that solves that exact problem, it absolutely is.
This is one of the biggest differences between classic paid search and AI discovery: ads (and visibility) won’t only map to the moment of direct intent. It will behave more like social: reach the right person based on relevance, even if the conversion happens later.
Agentic shopping: huge for some categories, contested for others
The last frontier here is agentic behavior: tools that don’t just answer questions, but execute actions, including purchases.
I’m bullish on agentic shopping for repeatable purchases. The value prop isn’t “better product.” It’s “you never have to think about this again.”
If an agent could just buy me new Converse every six months — same size, my favorite colors, pick one — I’m doing that in a heartbeat. I don’t want to fight filters, timeouts, and a parade of ads for shoes that look like Converse but aren’t Converse.
But there’s a real counterweight too: major platforms don’t want to give away the relationship or the data. Search data inside Amazon is incredibly valuable, and it’s hard to imagine them letting a third-party agent sit between them and the customer at scale.
So the likely future is a mix:
– platform-owned agents inside major ecosystems
– and agent-driven discovery that creates new opportunities for brands outside those ecosystems
Either way, the trajectory is clear: more of the journey becomes mediated by AI. Which brings us back to the start.
The takeaway
AI is a new front door to the internet.
It changes how people search, how they discover brands, and how influence happens — often without clicks. It will bring ads, but only if those ads can be integrated without breaking trust. It elevates the importance of inputs: content structure, third-party presence, and especially feeds for retail. And it forces a measurement reset — one that we as an industry are going to have to invent in real time.
The good news is we’ve done this before.
Not with the same tools, but with the same challenge: when the front door changes, the winners aren’t the ones who complain about it. They’re the ones who learn how to show up on the other side.
If you want to be discovered in AI search, start by being present where the models learn, refresh what you already have, and measure what matters, not what’s easy. Then iterate.
