I've spent my career on the media side of the table. I was one of the first ten employees at BuzzFeed and built its branded content business, then helped build Cheddar, then ran marketing at ClickUp. At ClickUp we had a real budget. We ran billboards, we had a very strong SEO program, and we did just about everything you can do in performance marketing. So I understand the instinct marketers have right now, which is to treat AI search as the next channel to optimize. Find the tricks, buy the tool, get the ranking.
I'd push back on that. What these tools reward looks a lot more like journalism than like SEO.
Your buyers already moved
G2 surveyed 1,076 B2B software buyers in March 2026 and found that 51% now start their research in an AI chatbot more often than in Google. A year earlier that figure was 29%. And the chatbot isn't just a starting point. According to G2, 69% of buyers changed which vendor they chose because of what an AI chatbot told them, and a third bought from a vendor they'd never heard of before.
Read that last part again if you run marketing at a company nobody has heard of yet. It means a buyer can go from never knowing you exist to signing a contract, and the thing that introduced you was an AI answer. It also means the reverse. You can lose a deal you never knew you were in.
Each tool picks sources differently
People talk about "AI search" as one thing. It's at least three, and they don't agree with each other much.
Google is the most straightforward. Its AI Overviews and AI Mode pull from the same index as regular search, and Google's documentation says "There are no additional requirements to appear in AI Overviews or AI Mode." You don't need special AI files or new markup. Google's newer optimization guide keeps coming back to content that's valuable and unique rather than generic. If you already rank, you're in decent shape here.
ChatGPT changed the game in May. Starting May 7, it began putting clickable brand names directly inside its answers. Similarweb found that ChatGPT referrals jumped 157.7% week over week afterward, and homepage referrals rose 354.7%. The share of ChatGPT traffic landing on brand homepages went from roughly a quarter or a third to about 60%, and it has stayed there. ChatGPT is now introducing brands to buyers, which is a very different job from sending readers to an article.
Perplexity searches the live web for almost every question and shows its sources openly. It also picks different sources than ChatGPT does. A 2026 citation report from 5W estimated that only about 11% of domains get cited by both, and an Ahrefs study of 15,000 queries found only 12% of AI-cited URLs also showed up in Google's top 10.
So if your plan is "rank on Google and the rest will follow," the data says it won't.
Most of what AI says about you isn't on your website
This is the part I think B2B marketers underestimate. A Semrush study found Reddit and LinkedIn among the top five most-cited domains on ChatGPT, Google AI Mode, and Perplexity. G2's 2026 Buyer Behavior Report found review sites have now passed AI chatbots as the top influence on which vendors make a buyer's shortlist, 38% to 37%.
In other words, when a buyer asks an AI tool about your category, the answer is being built largely from what other people say. Practitioners on LinkedIn. Users on Reddit. Reviewers on G2. Editors at industry publications.
And those sources move around. The same Semrush study showed ChatGPT citing Reddit in nearly 60% of responses in early August 2025, then around 10% by mid-September. Anyone who tells you they've figured out a permanent formula for one platform is going to have a rough quarter at some point.
Why a publication beats a playbook
When we started Outlever, the idea came straight out of what I'd seen in news. Media companies were producing branded news for advertisers, but the brands didn't own it and couldn't distribute it. So we flipped it. We build news publications that companies own outright. I've described it as news publishers in a box, and that's still the simplest way I know to say it.
What I didn't fully appreciate at the start was how well that model would line up with how AI tools choose sources. Think about what a real publication produces. Original reporting. Named experts with titles and companies. Data with a date and a source attached. Coverage of what's happening this week in a specific industry. That's the material every one of these tools is hunting for, because it's the material they can't get from the thousand other pages saying the same thing.
The interviews matter most. Across our clients we've done thousands of them, and every one creates something that exists nowhere else on the internet. Then the person we interviewed shares it on LinkedIn, and their colleagues comment, and sometimes it gets picked up by a newsletter or another outlet. So the brand ends up with an original source on its own site and a trail of mentions on exactly the platforms AI pulls from. One interview does both jobs.
That's what I meant when I told the Superpath community, in an AMA this June, that we're building the source, not racing to be first to a story. Being the source is what gets you cited.
What I'd tell any B2B marketing team to do
Check your plumbing first, because it's boring and people skip it. Make sure your robots.txt file and your CDN or hosting setup aren't blocking the crawlers these tools use. Google uses Googlebot. OpenAI and Perplexity run their own crawlers, called OAI-SearchBot and PerplexityBot. Some security tools block them by default, and I've been surprised how many companies don't know.
Then stop publishing things that already exist. If your article could be swapped with any of the top ten results and nobody would notice, an AI tool has no reason to pick yours. Publish your own numbers, your own interviews, your own read on what's changing in your market.
Write the way a good reporter writes. Put the answer high up. Define terms in plain language. When you use a number, say where it came from and when, in the same sentence. AI tools tend to lift a single paragraph out of a page, so each paragraph has to make sense on its own. One Q2 2026 analysis found that pages built around quotable data, clear definitions, and tables were cited more often than higher-authority pages without them. That matches everything I learned in newsrooms. Clear writing travels.
Find out what your buyers are really asking. Your sales team hears the questions every day. Type them into ChatGPT, Perplexity, and Google yourself and look at who gets named and which sources get cited. That list is your editorial calendar.
Take care of your homepage. Since ChatGPT started sending most of its referrals there, your homepage is often the first thing a buyer sees after an AI tool mentions you. It should say what you do, who it's for, and why you're different, in words a stranger understands. Keep your facts consistent across your site, LinkedIn, and review profiles too. When those disagree, AI tools get confused, and they'll describe you wrong.
Get your people talking in public. Executives posting under their own names on LinkedIn, showing up on industry podcasts, being quoted in other publications. Keep review profiles current. All of it feeds the answers.
And please don't try to fake it. Google has said that chasing inauthentic mentions to shape AI answers isn't likely to help, since its AI features use the same safeguards as regular search, as Semrush reported. Planted Reddit comments get found out, and the damage to your reputation lasts longer than any bump in visibility.
How we think about measuring it
There's no perfect tool yet, and I'd be skeptical of any vendor who says otherwise. What we do is simple. Once a month, run a fixed set of real buyer questions through ChatGPT, Perplexity, and Google, and track whether the brand shows up, how it's described, and which sources get cited. Google reports AI Overviews and AI Mode traffic inside the regular Performance report in Search Console. Watch referrals from chatgpt.com and perplexity.ai in analytics. Look at the trend over several months rather than any single week, because these platforms change their behavior constantly.
The number I care about most isn't on any dashboard, though. It's whether the right buyers know who you are before your sales team ever calls them. AI answers are becoming one of the main ways that happens.
Questions marketers ask me about this
Do we need an llms.txt file or special schema?
Not for Google. Its documentation says you don't need new machine-readable files, AI text files, or special structured data for AI Overviews or AI Mode. Standard structured data is still worth having for regular search.
If we rank on Google, will ChatGPT cite us?
Not reliably. Ahrefs found only 12% of AI-cited URLs also appear in Google's top 10. Good SEO helps with AI Overviews, but you have to check ChatGPT and Perplexity separately.
Why does Reddit show up so much?
Because it's full of real people describing real experiences, and AI tools treat that as evidence. For B2B brands, it means what practitioners say about you in public can matter as much as your own website.
How long before we see results?
Pages that already rank can show up in AI Overviews quickly. Becoming a brand that gets named consistently across all three tools usually takes months of steady publishing and real mentions elsewhere.
Can a small company compete with the big brands?
Yes, and I'd argue small companies have an edge on specific questions. AI tools answer narrow questions all day, and a focused company with original data or strong expert interviews on its topic can be the best source available. G2's finding that a third of buyers bought from a vendor they'd never heard of tells you these answers aren't reserved for the biggest names.
Where I land
I don't think AI search needs a brand new discipline. I think it needs companies to finally do what media companies have always done, which is cover a subject so well that people trust you on it. The tools are just making that trust visible. The brands that treat this as a hunt for tricks will spend the next few years reacting to every platform change. The ones that build a real source will keep getting cited, because they'll keep having something worth citing.
The State of Brand is published by Outlever, which builds and operates owned media publications for brands.





