Brittany Trafis on AI Search as a New Channel Beyond SEO
BEST QUOTES
“AI search is not SEO 2.0. You have to treat AI Search as that new channel and think of your strategy differently.”
— Brittany Trafis
“Quit treating AI search as SEO. If you keep treating it like just keywords and rankings, you will get keywords and rankings in Google — but you won’t show up in ChatGPT, Claude, and Perplexity.”
— Brittany Trafis
“AI Search put us all at the starting line together. Small businesses can move a lot faster than large organizations — just because you’re small, don’t put yourself out of the race.”
— Brittany Trafis
JEREMY RIVERA
Hello, I’m Jeremy Rivera, your Unscripted SEO Podcast host. I’m here with Brittany Trafis of Soarion Digital, who’s going to introduce herself and focus on the things she has done previously that built the expertise, experience, authority, and trust that should make us trust what she says about digital marketing.
BRITTANY TRAFIS
Thank you for having me, Jeremy. It’s great to be here on the show today.
JEREMY RIVERA
My pleasure. So where have you been and what did it teach you as you kind of jump into this digital marketing world?
BRITTANY TRAFIS
Yeah, so I have always been involved in marketing in some form — whether it was as a graphic artist making flyers for a food safety distributor, or the intern running and setting up conference booths. And as I look back, it even starts when my first job was working at Chi Chi’s as a hostess, helping greet and seat people before there was technology to tell you just what table to go put them at.
That all moved me into really understanding graphic design, trade show booth design, and eventually the difference between branding, lead generation, and awareness — phrases you learn in college but don’t truly understand until you’re in it. Where it really all came together for me was understanding the startup world.
Going from corporate to startup, I worked at a company growing their own email platform in the early-to-mid 2000s — a time when ESPs were just blowing up everywhere, honestly similar to people in AI today. Working hands-on there started to move into understanding customer engagement, account management, strategy, and measurement — I could track across websites. All of these elements really started to build together. The last 10 years I spent helping lead a leading digital marketing agency, and then AI hit.
As I looked at AI hitting and where we’re at today, it all came together. My passions, I didn’t realize it until these two letters really came into the world. It comes back to solving problems and seeing what that gap is and helping people fill that gap. I always compare it to firefighters — the house is on fire, people are running out, and they’re running in. Very few people want to run in and dive head-first to fill that gap. That’s where my experience in technology, marketing, digital, and customer relationships has really built me for this moment.
Soarion Digital’s USP: AI-Native Marketing Agency
JEREMY RIVERA
So for Soarion Digital — what’s your USP? Somebody steps into that elevator, you’ve got 60 seconds. This is what we are, this is what we do, this is what makes us unique. Go.
BRITTANY TRAFIS
Soarion Digital is an AI-native marketing agency and we help companies adapt to the shift in search engines. Search has changed and we’re here to help you show up in AI search engines — in generated results — not just in traditional search like Google.
JEREMY RIVERA
What needs to adapt from a fully traditional SEO stack to add reality for these new systems? Let’s say the small business understands: we’re writing four blog posts and building four to eight backlinks as our SEO program. How does that change or adapt within an AI-native agency?
BRITTANY TRAFIS
Lots to unpack there. The number one piece I say is AI search is not SEO 2.0. SEO as a foundation is still important — your backlinks, your keywords, that strategy still exists and is foundational for AI search. Your technical components, keywords, and content creation still matter.
But what has really changed are the important elements within your technical website setup. Things that were so important 20 years ago from an SEO perspective — schema, meta tags, descriptions — started to get quieter, and now adjusting your schema can really improve your citation rates very quickly within AI search.
You still need SEO, but you have to treat AI Search as a new channel and think of your strategy differently. It would almost be like when Facebook Ads came out — if you tried to run them like Google Ads, it wouldn’t work. That’s the same relationship between AI Search and SEO.
Schema, Technical Elements, and How LLMs Read Your Site
JEREMY RIVERA
I’m curious — optimization in SEO has turned specific corners where technical perfection of implementation would matter a lot. If you added schema but it didn’t validate properly, you might as well have not done it. But with LLMs, I get the feeling it’s almost not about the schema being technically perfect — it’s more about a secondary document clarifying what things are about, in a squishy, LLM-friendly way rather than rigid programmatic compliance.
BRITTANY TRAFIS
Exactly. Google is very rigid — they’ve trained us for years. How long a title tag — that’s set by Google and we follow it. But what the LLM cares about is what is that schema telling me? Take FAQ schema — AIs love it. It’s like, you’re going to tell me the question and the answer, and I can just grab that without going any further. Because they’re trying to move very quickly compared to Google doing more of the rankings.
The LLMs are looking for the right response to the question they have, and they’re moving fast. That’s why that technical element is so important. Now, if your schema is not set up correctly and it’s timing out, then it’s going to move on and ignore you. But it is going to be looking for that surface-level content first.
Training Data vs. Retrieval: Building Brand Visibility in Both
JEREMY RIVERA
One of my favorite interviews was with Jason Barnard of Kalicube, an entity-first optimization expert. He described a useful framework: there’s the training dataset, and then there’s the retrieval process — if it’s not in that dataset, they’ll use Google searches as a fancy wrapper to augment. Which of those two places is it most important to start building brand visibility?
BRITTANY TRAFIS
I agree with that structure — that’s absolutely what’s happening. There’s a new industry stat that 50% of content being served within LLMs is less than 13 weeks old. So you want to get into the overall knowledge base it’s trained on, but from a retrieval perspective, you have to be showing up there too, because there’s only so much it will maintain. You really do have to be in both.
What I can tell you is we’re monitoring clients’ websites and we’re seeing agents crawl them within hours of something new being published — where Google you’re still waiting six to eight weeks for them to even index and recognize your changes. If you want to move quickly and see results fast, that’s where you’ll see it on the AI side.
AI-Native Workflows, Innovation, and Consumer Behavior Shifts
JEREMY RIVERA
So, adversity breeds innovation, you know, like people were concerned about, you know, AI data centers chewing up power. And so they’ve now got like behind the meter LNG power, you know, that’s off the grid, you know, to fill on the backend. What are some of the innovations that you’re seeing in the industry or adaptations that you’re seeing that are a response to the challenge brought to the market by AI?
BRITTANY TRAFIS
I’m here in Ohio and the data centers are just popping up like wild here. In terms of innovation — for Soarion Digital, I ran traditional agencies and just saw how much change AI was going to bring. I took a step back and thought: what’s this look like if I build AI first and people next? There’s still room for people, we still need the expertise. But it completely changed our processes and workflows. Workflows that used to have 16 to 25 steps now have like four — because you have to reimagine and open your mindset.
The people leading are those who stop and think what should it be instead of what has it always been. And from the consumer side — did you Google today or did you Claude or ChatGPT today? Consumer behavior has just completely changed. I saw that Ask Jeeves officially sunset today, and I made a joke that Jeeves had a grandson named Claude, because I just ask Claude everything now.
JEREMY RIVERA
I asked Claude: ‘What would be a good way to add internal links?’ It suggested a related-article callout to the previous episode, categorizing articles by primary themes, building a hub page, and cross-linking. Then I added the Royal MCP plugin for WordPress, so now I can push updates directly from Claude Desktop. What would have been at least a week of hands-on work was done in a session. If you take your traditional tooling and find a way to connect it, there’s so much more creativity you can apply without the restriction of ‘that would take so long.’
BRITTANY TRAFIS
The hardest part is making the time to figure out how you can relieve the time for yourself. I have a long list of things I want to test. But when you make that first step and go have the conversation with Claude, once you’re in there, you start to quickly adapt. It can be freeing — you spend more time on the human elements you like and less on the tedious, mundane pieces. I use an example: I’m really bad at managing my inbox. But Claude now every morning sends me a digest — it’s managing my whole inbox for me. It’s like having a personal assistant.
JEREMY RIVERA
I dipped my toe into getting a virtual assistant to help with podcast post-processing, and I realized it didn’t work because I hadn’t first worked out all the steps with Claude myself. My last interview with Mason McCumber, an AI automation specialist, drove this home — he said: ‘You keep putting in prompts but you need to make these as skills in Claude so they’re repeatable.’ Now you can stack your skills and have agents execute skills in a sequence. It totally saved me 20 minutes per podcast interview. You can read more about how AI tools are reshaping SEO workflows in my interview with Jason Berkowitz on the SEO Arcade blog.
BRITTANY TRAFIS
Absolutely. A lot of this now we can train ourselves, so it knows what I’m looking for or what my style is. As you give it those specific skills, it stays more trained. The time it takes starts to reduce — take the time to make the time to relieve the time — but each of those components are things we see people starting to prompt and have conversations with AI on.
Data, Measurement, and the LLM Black Box
JEREMY RIVERA
Google about a year ago messed with the rank tracker industry by removing 100-results-at-a-time queries, and impressions dramatically dropped in Search Console. Are we looking at a similar problem when it comes to tools trying to get information about brand visibility in AI? Is it just a snake eating its own tail — tools asking queries to build databases?
BRITTANY TRAFIS
It is the wild west right now. We are watching this channel be born, and we’re still trying to settle on a name — AEO, GEO, AIO, AI Search — we haven’t settled yet.
The biggest challenge right now is you can’t get the data out. We’ve been trained by Google, HubSpot, and all these different marketing tools — they give us the data. We know who clicked, where we were served. That’s not coming out of any of these LLMs right now. They’re a black box.
Some monitoring tools say they’re getting data, but really it’s just from prompting the LLMs to see what comes back. Always, huge asterisk: these tools are directional. Do not think this data is Google Search Console level data. The LLMs are also so highly personalized — if you and I did a search at the same time right now, we would get two completely different results. You need the monitoring data as a signal for your strategy, not as hardcore data.
What I always encourage is to look at your web analytics as your final source of truth — what was that AI referral traffic to your site? You can see that. That means they actually found you and hit your page. That’s the soundest data point you can get right now.
JEREMY RIVERA
I did just see that Copilot added data in Bing Webmaster Tools, with more metrics announced. Unfortunately, for projects like a home builder in Cookeville, Tennessee, they’re not really giving useful LLM-based data. But I’m starting a new project with a national brand and I’m curious — they’ve always been more transparent with their data than Google. If anybody is going to give us a dashboard, it’s likely to be Copilot and Bing.
BRITTANY TRAFIS
From what I’ve read, the ChatGPT ad pilot was limited to nationally recognized top brands, and they got limited data — nothing like going into a dashboard to see audiences and manipulate who you’re serving to. What I did read last week is they’re starting to put in something similar to the Meta Pixel — a new pixel that can be put on websites to track their ads. So they were able to get limited data on people clicking through to the site, but it stayed there based on what I’ve been able to read and hear.
Hottest Take: Stop Treating AI Search Like SEO
JEREMY RIVERA
I’m going to give you a moment to stand on a soapbox and shout at business owners and prospective clients about the biggest mistake they’re committing. What is your hottest take?
BRITTANY TRAFIS
Quit treating AI search as SEO. Stop. If you keep treating it like just keywords and rankings, you will get keywords and rankings — in Google. If you’re trying to show up in ChatGPT and Claude and Perplexity, where people are more engaged and want that information, you’re not going to see it. Read more about why AI Search is not SEO 2.0 on the Soarion Digital blog.
And if your SEO agency says ‘we got it, we’re doing AEO, we’re doing AI search’ — really lean in, ask three more questions deeper. I think you’ll find out that they’re just treating it like SEO.
JEREMY RIVERA
What do you wish small business owners knew about AI visibility that we haven’t covered yet?
BRITTANY TRAFIS
I love AI Search for small businesses because it is a new channel. It put us all at the starting line together. How many times have you worked with a small business that says, ‘I’m never going to get that first-page ranking, I don’t have millions to throw at ads, I can’t beat my large competitors.’ Guess where you do stand a chance? In AI Search.
We are working with brands who have a quarter of the budget of their big competitors in paid advertising, but now within AI search, they’re starting to show up and have a higher presence than those larger brands. And it’s because they’re putting a strategy behind it. Small businesses can move a lot faster than large organizations. So just because you’re small, don’t put yourself out of the race. You’re actually already in the race and could be ahead of the larger guys.
AI Content Strategy for ‘Boring’ Industries
JEREMY RIVERA
So I guess that helps if you’re in a boring industry, because I was talking with Matt Brooks of SEOteric about how do you deal with, you know, marketing for a company that doesn’t do sexy stuff, you know, they install concrete walls. That’s as interesting as washing windows. But how do you approach the interest level angle? If somebody is very interested in putting that concrete fence in their backyard, then it doesn’t matter if it’s not a Maserati, because it’s personally important to you. What are your steps to leverage AI to craft deeper responses, better content, more fulfilling content for your users when the subject matter tends to be a little more dry?
BRITTANY TRAFIS
Yeah, absolutely. There’s a few components there. So yes, it’s boring. And so it’s not going to be this beautiful ad or anything. But within AI search, to your point, the person who needs to put the brick wall or cement wall in their backyard, they care. And they’re coming in and they might not even know you exist. And they’re starting to have that conversation. So what we always encourage our clients and others is really start thinking about what that longer-form structured content looks like. Break it down. What are those structured components within that content? The definitions — that’s the first piece — comparisons, pros and cons, and then point of view. Go through the thought process of somebody who might have the problem and is now starting to solve it, and have multiple pieces of content that can fit there, but have them structured very cleanly so that the AI can pull that and then serve you as that answer.
Go through the thought process of somebody who has the problem and is now starting to solve it. Have multiple pieces of content that can fit there, but have them structured very cleanly so the AI can pull that and serve you as the answer. As they’re going through that content and seeing your name and building that trust, you’ve built credibility with them before they’ve even clicked on your website — because you’ve met them where that content needs to be.
Designing Human + AI Workflows: Avoiding the Hive Mind
JEREMY RIVERA
I just published my interview with Nick Eubanks, and we were talking about how some people are using AI capabilities in the wrong place in their workflow — putting AI in charge of creating the content brief instead of fulfilling it. There’s also the ‘hive mind’ problem: creative writing prompts that for humans create vastly different stories, but when fed to different AI models, all came up with surprisingly similar stories. What tooling or processes do you add to make sure that as you integrate AI you’re not producing same-y content? You can also see how entity optimization and brand signals play into this from my interview with Mark Williams-Cook.
BRITTANY TRAFIS
I am a firm believer that some of our greatest strengths as human marketers are in creativity, strategy, and domain expertise. Your domain expertise really informs how you think creatively and strategically. In terms of how we work: the human understands what you need as a business and digests those components. Then we have the AI look at the data and give us insights. Then the human pieces that together to form the overall strategy — which the AI can then start to write content on — and then the human is in the loop at the end reviewing it.
Really lay out what your workflow is and identify: where do you need that human strength? Where do you actually need human creativity, strategy, and domain knowledge? Plug them in and don’t move that step. That step is locked in — as easy as it might be to put AI in that place, do not lose those human-centric pieces. Let AI automate the other pieces, and have a human get in at the end to make final changes.
Closing: Where to Find Brittany & The Open Question
JEREMY RIVERA
As a wrap up — tell the people where they can find you, connect with your company, and if you’re on a particular social media channel. Also if you have anything interesting coming up.
BRITTANY TRAFIS
You can find us at soariondigital.com. I’m also pretty active on LinkedIn — you can follow me, Brittany Trafis, or Soarion Digital on LinkedIn. We have quite a few events coming up: I’m going to be at CMO Huddles doing a Midwest tour meeting with top CMOs talking about this exact topic. We’ll also be at the Marketing AI Conference in October here in Cleveland, where thousands of marketers come together to talk about AI. If you have any questions, reach out on LinkedIn or soariondigital.com. I’m happy to help you identify how you move from just using SEO to AI Search as a new channel.
JEREMY RIVERA
What pain point or problem do you want to float to the industry and ask the next SEO professional that comes on for their opinion?
BRITTANY TRAFIS
The biggest pain point I have right now is how to keep up with it all. Things are changing so quickly — I bet while we had this conversation, something rolled out. How do you keep up with everything happening in AI, and how do you filter what you need to focus on versus what’s just noise? There’s a lot of noise in the news right now around AI. If anyone has figured out a trick, I would love to know.
JEREMY RIVERA
I’ll make sure to get back to you after the next interview. In the meantime, if anybody wants to find Brittany, I’ll make sure those resources she mentioned are linked in the show notes. Thanks for stopping by.
BRITTANY TRAFIS
Absolutely, thanks for having me.
Key Takeaways
- AI Search is a new channel, not SEO 2.0. Stop treating it like keyword rankings and start treating it as distinct — just like Facebook Ads required a completely different strategy from Google Ads.
- Schema still matters, but for different reasons. LLMs love FAQ schema because they can grab question-and-answer pairs instantly. The technical element is still critical, but the goal is semantic clarity for a fast-moving crawler, not rigid structural compliance.
- You need visibility in both the training data and the retrieval layer. Approximately 50% of LLM-served content is less than 13 weeks old, so fresh, crawlable, well-structured content matters for retrieval, not just long-term brand authority.
- For ‘boring’ niches, structured content is your competitive advantage. Break content into definitions, comparisons, pros and cons, and point-of-view pieces. The AI serves you as the answer when your content is clean and structured.
- AI Search has leveled the playing field for small businesses. Brands with a fraction of larger competitors’ budgets are achieving higher AI-search presence through deliberate strategy and the agility advantage that comes with being smaller.
- Human creativity, strategy, and domain expertise must be locked into your workflow — do not let AI replace those steps. Use AI for data analysis, content execution, and automation; keep humans in the loop for strategy creation, brief development, and final review.