[0:00] Rich Smith: Welcome to Revenue Science. I'm Rich Smith, CMO, founder, investor, executive adviser, and lucky enough to be your host. Today's episode is a little different — we're diving into one of the biggest shifts happening in B2B growth right now: how AI assistants like ChatGPT, Perplexity, and Google AI are changing who gets discovered, trusted, and ultimately purchased from.
Joining me today is Ray Young, AI search strategist, researcher, and president of Razor Sharp PR. We're going to unpack what it really takes to win AI citations, why earned media and substance matter more than ever, and what CEOs and CMOs should do before the competitive gap compounds.
[1:23] Rich Smith: Ray, welcome to Revenue Science. Can you introduce yourself and tell the audience about what you do?
[1:34] Ray Young: I'm the president and founder of Razor Sharp PR. I've been in the PR and media relations business for almost 30 years — it's changed dramatically over that time. I started Razor Sharp PR in 2011 after leaving another agency because I wanted to lean hard into social media as a communication channel. Our focus is helping businesses, CEOs, and organizations raise their visibility with investors, customers, clients, and any stakeholder that matters to them — making sure the right messages reach the right people, especially those who might misunderstand what you do.
[2:37] Rich Smith: Why should a CEO care about AI search specifically?
[3:34] Ray Young: Because AI is deciding who gets recommended and who gets ignored. Google's AI overview handles roughly two billion interactions a day. ChatGPT processes around two and a half billion prompts. In B2B, buyer adoption of AI assistants across the entire purchase process is approaching 100%. If your company isn't showing up in AI search, it's not just a visibility problem — it's going to become a revenue problem.
[4:38] Rich Smith: Is this a niche trend or has it already changed how buyers find companies?
[4:43] Ray Young: It's already changed, and it's changing rapidly. AI traffic in the US grew 10 to 12 times between mid-2024 and early 2025, and those visitors convert at four to ten times the rate of traditional organic search traffic — that's 400 to 900% better conversion. At the same time, when Google puts an AI answer at the top of the page, click-through rates on everything below it, organic and paid, drop significantly. The old channel is shrinking while the new one is producing far better buyers. Google just announced they're expanding their answer box for the first time in roughly 25 years, essentially because it's all moving toward AI.
[6:09] Rich Smith: Why do AI-referred visitors convert at higher rates?
[6:40] Ray Young: It's all about intent. A traditional Google search might be someone casually browsing or doing early homework. When someone uses ChatGPT, Grok, or Claude, they're asking a very specific question and the assistant returns a very specific recommendation. The AI does a much better job of filtering, so when a potential client lands on your website they're already well down the road toward a buying decision. ChatGPT in particular skews toward deeper evaluation and decision-stage queries, versus the casual browsing that traditional search captures.
[9:38] Rich Smith: It feels like a trusted friend making a recommendation — and that removes the uncertainty that's one of the biggest barriers to buying, which is why conversion goes up.
[10:18] Ray Young: Exactly. And the way people use their AI assistants is fascinating — some name them, some use them almost like a psychologist to think through decisions. The more you use these platforms, the more they start to know what you already own, what you've bought before, and can recommend something that fits your specific situation. That deepens the trusted-friend dynamic considerably.
[11:29] Rich Smith: What does being invisible to AI actually look like?
[11:36] Ray Young: If you ask ChatGPT or Grok a relevant question and your company doesn't come up but your competitors do, you're invisible. It usually comes down to one of two problems: either the right content doesn't exist, or it exists but AI can't reach it. We worked with a company that had a genuinely strong community, but all of it lived in Discord and GitHub — formats AI crawlers can't search. The fix wasn't to create new content; it was to make their existing content crawlable.
[13:00] Rich Smith: How do AI assistants actually decide what to recommend?
[13:05] Ray Young: It comes from two places. First, memory — everything the AI learned during training, its existing worldview. Second, real-time search — the AI going out to the internet to find answers before responding to your query. Those two get blended together and cross-checked for accuracy, which is also why AI sometimes gives wrong answers: the training data was incomplete or the right source wasn't findable. You want to be in the AI's long-term memory and in the sources it retrieves in real time. Miss both and you're invisible.
[14:17] Rich Smith: You've written about something called the parametric knowledge flywheel. What is that?
[14:30] Ray Young: Here's how it works: if your content gets picked up in real-time retrieval today — say a CEO interview on Bloomberg that gets archived on YouTube — it gets stored in the AI's memory. Research suggests that once a brand is in the AI's memory, it's much easier for the AI to cite that brand again because it's already the starting point of the answer. That's the compounding part: getting into memory makes future citations more likely, which creates more training data, which makes citations even more likely. Getting into that memory early is crucial.
[15:31] Rich Smith: Why is third-party coverage dominant in what AI cites?
[15:36] Ray Young: Because AI wants to give you the best, most credible answer it can. It will pull some things from your own website and blog, but a large portion comes from third-party sources — Wikipedia, news sites, Bloomberg clips on YouTube, contributed articles on MarketWatch, Wall Street Journal stories. Those rank very high on credibility and AI naturally gravitates toward the most credible sources it can find. A no-name blog with three readers isn't getting cited. What this comes down to is the basic principle we've known in PR for decades: what someone else says about you is far more credible than what you say about yourself.
[19:14] Rich Smith: There's also a recency bias in AI citations — how significant is that?
[19:20] Ray Young: Very significant. About 65% of sources cited by AI were published in the past year, and about 80% in the past two years. Google's Gemini has the strongest recency bias, with roughly 85% of citations coming from the last three years. AI basically accelerates content decay — if your content is more than a year old, in most cases it's not getting found. That said, a highly specific, well-researched white paper from years ago could still surface if someone does a very targeted search. The practical takeaway: publish quality content more consistently and keep it recent.
[20:56] Rich Smith: Do these platforms filter out content that adds nothing new?
[21:01] Ray Young: Yes. Substance beats style. If you're just regurgitating what a hundred other sources have already said, AI will filter you out. What gets cited is original data, proprietary research, unique surveys, expert analysis, and perspectives that are hard to find anywhere else. AI systems are looking for content they actually need. If you can provide a perspective that's rare or data nobody else has published, that's what gets surfaced. "AI slop" — content that technically exists but says nothing new — doesn't make the cut.
[22:20] Ray Young: That's not to say repurposing content is a bad strategy — it can work well. But you have to update the point of view, layer on new data, tie it to something current, or invert it entirely. "Five things you should know about investing" becomes "Five mistakes to avoid in investing." Same core, freshened up, different enough to be useful.
[24:05] Rich Smith: You've said brands that settle disputes get cited. What does that mean?
[24:11] Ray Young: AI isn't great at synthesizing genuinely conflicting viewpoints — it tends to default to whatever the majority says or whatever it already believes. If your content provides a well-supported, evidence-backed resolution to an open question in your industry, the AI can use that as a tiebreaker. This pattern is especially strong in B2B, where comparison-based content gets cited at a much higher rate than in B2C. If your CEO has 30 years in the field and produces content that definitively resolves a question nobody else has answered cleanly — backed by data — you're in a strong position to get cited.
[25:49] Rich Smith: Is winning AI citations really just about producing more content and increasing volume?
[25:57] Ray Young: Absolutely not. Volume is good, but it's not the driver. One well-polished piece with proprietary data and a genuinely interesting point of view will get cited more than a hundred pieces of AI-generated filler. AI assistants actively resist shallow, regurgitated content — they want a real point of view. It could even mean producing less but better quality. The AI slop problem is real: there are people publishing AI-generated books and blog posts en masse, and the platforms are increasingly good at filtering that out. Quality is what sets you apart.
[27:59] Rich Smith: The other key I'd add is figuring out what you're uniquely qualified to talk about, in a way that also supports the credibility of what you do. Pick a lane, stay in your swim lane, and build a content calendar around consistent points of view in that space.
[28:54] Ray Young: A hundred percent. Media won't cover you if you're bouncing between unrelated topics. I use a Venn diagram with clients: what are you qualified to talk about, what moves the needle for your business, and what does your audience actually care about? Where those three overlap — that's where you focus.
[30:04] Rich Smith: What are the broader changes happening with Google search and what's the impact?
[30:25] Ray Young: The direction is 100% AI, eventually. Google is already pushing everyone toward AI-generated answers, and they'll figure out how to run paid ads inside AI responses. For now, the answer box is expanding to be essentially all AI. I personally started using ChatGPT about a year ago and it cut my Google searches in half overnight — now with Claude and other platforms my traditional Google usage is down to maybe 10-20%. In B2B, especially in industries like finance or real estate where buyers are thoughtful and deliberate, AI adoption in the research and purchase process is already near 100%.
[33:24] Ray Young: One thing I always do with new clients is ask: what are your buyers actually asking? What questions surface over and over in sales conversations? There are always five, ten, fifteen questions that come up constantly. Build content that answers those questions — that's what AI will find when your customers and prospects go looking.
[34:53] Rich Smith: If I had to pick two content priorities, it's third-party earned media coverage and first-party content that helps your buyer navigate a problem — not chest-pounding about how great you are. That self-promotional content gets discounted.
[35:18] Ray Young: Exactly. Think about it personally: if you know someone who constantly brags about themselves, do you trust them? But if four different people independently tell you that person is the authority on something, you trust that completely. AI works the same way — it goes to trusted third-party sources, not your own homepage.
[37:02] Rich Smith: How would you advise a CEO to test whether AI is actually helping or hurting their business right now?
[37:08] Ray Young: Go put your name and your company into ChatGPT, Google's AI search, and a couple of others and see what comes up. Are you there? Are competitors showing up instead? Then go into Google Analytics — they now have an AI-referred traffic section — and see where your traffic is actually coming from. If you have historical data, compare it and look at the trend. Chances are your traditional organic search volume has declined or at least plateaued, and if you're not making up the gap with AI-referred traffic, you're beginning to fall behind.
[39:27] Ray Young: Every time a competitor gets cited, that citation feeds their flywheel — it gets archived, it becomes training data, it makes the next citation more likely. If your customers aren't training AI systems on who you are and what problems you solve, you're going to get left behind. And unlike traditional SEO where you could theoretically spend $40K on paid ads and catch up in a quarter, AI is a longer compounding game. You can't buy your way into the AI's memory overnight.
[43:08] Rich Smith: One real danger is confirmation bias — these platforms tend to tell you what they think you want to hear. I notice it in my own use and I start to question the platform's credibility when it's constantly validating every idea I have.
[44:14] Ray Young: That's why you have to prompt it correctly. Tell it explicitly: don't tell me what I want to hear, answer me honestly from an outside perspective. You have to keep it honest or it will just reinforce whatever you already believe, which is exactly the opposite of what you need it for.
[45:48] Rich Smith: If you had to give a CEO one step to take this quarter, what would it be?
[45:53] Ray Young: Start testing. Build a set of 100 to 200 questions your buyers actually ask — talk to your salespeople, customer success team, anyone who interacts with clients regularly. Run those questions through the major AI assistants and record who gets cited. Pull your clickstream data to see where your buyers are actually spending time and reading. Get a snapshot of your current visibility so you can make an informed decision about whether you have a gap to close and how urgent it is. Some audiences have near-100% AI adoption already; others have lower adoption and you may have more runway. The data will tell you which situation you're in.
[48:32] Rich Smith: Where can people learn more about you and connect with you?
[48:44] Ray Young: Our website is razorsharpr.com — you can find our blog posts, contact information, and everything else there. You can also email me directly at ray@razorsharpr.com.
[49:06] Rich Smith: We'll put all of that in the show notes. Ray, thanks so much for being here — I think our audience is going to get a lot out of this one.
