Growth Unhinged is proudly supported by AirOps
AI agents are quickly rewriting the rules of growth. On May 13th in NYC, the AirOps Next Conf brings together the leaders building the future of AI-driven marketing, search, and content systems.
Expect fresh research on AI search, hands-on agent workshops, real playbooks, and live demos. Speakers include growth leaders from Anthropic, Harvey, Ramp, HelloFresh, and more. If you care about AEO, agentic workflows, or where growth is headed next, this is one worth paying attention to. Apply for a spot here.
When I started writing about AI search last year, it felt like something that was coming soon yet hadn’t quite arrived. It wasn’t yet a meaningful channel for the average startup, and you could kinda wait and see before putting big dollars into it.
Well, that’s no longer the case. Answer engine optimization (AEO) is the single biggest channel where B2B marketers are increasing investments this year. AI searches probably represent at least one-third of all searches these days when you combine AI Overviews (21% of Google keywords as of November 2025) and LLM queries (ChatGPT alone does about one-fifth of Google’s query volume).
The BFD: AI responses recommend products rather than simply provide a list of links. This means the buyer increasingly generates a shortlist of products (or even makes a buying decision) before ever visiting your website. And AI engines are becoming the main way that people interface with your product in general (see: Salesforce going ‘headless’).
With AEO maturing rapidly, I reached out to more than a dozen marketers about what they’re doing differently and what’s working right now in AI search. We covered AEO metrics, how to get AI to say the right things about your products, and why YouTube (not Reddit) is the latest AEO battleground.
1. Shift from AI visibility to influenced pipeline
AI engines don’t send much referral traffic, and instead keep users within their walled garden. Last year the goal was visibility: share-of-voice in AI queries became the dominant metric.
While visibility still matters (that’s the starting point), it’s not enough. Marketers are shifting to comprehension and pipeline influence. I think this reflects the maturity of the category and helps justify growing AEO budgets.
Comprehension: How well AI engines describe your products relative to competitors
Pipeline influence: How much pipeline can be attributed to AI engines whether directly (referral traffic) or indirectly (self-reported or manual attribution)
beehiiv, for example, shifted to measuring LLM-attributed signups according to CMO Darren Chait. With this shift came a greater focus on getting LLMs to better describe beehiiv and its differentiation.
“This year we have shifted from measuring LLM visibility to LLM-attributed sign ups. Last year we prioritized the typical AEO optimizations, and launched a successful Reddit strategy, which led to material growth in how often beehiiv was included or cited by the models. But in many cases, the LLMs were not doing a good job of describing beehiiv, its differentiation, and where it sits in the market. We started re-writing/re-structuring the pages that the models were citing, like our enterprise landing page, and we soon saw significant growth in leads and signups attributed to LLMs – with no real change in visibility metrics.”
2. Solve the consistency problem
Many marketers noticed that LLMs weren’t doing a good job of describing their products. While it’s easy to blame LLMs for this, the reality is that most brands were making it really difficult for AI. (“It’s me, hi 👋, I’m the problem. It’s me.”)
Reply realized that they had over 500 blog posts that described the product differently. Their solution was to align around a single idea and make sure that was reflected in all key pages.
“One of our biggest realizations at Reply was that we had 500+ blog posts and 1,000+ mentions all describing us differently. We aligned everything around one idea -- sales engagement and AI SDR -- and made that the consistent focus across key pages and mentions. That alone got us into the top 3 and helped us show up in ~25% of 300+ tracked prompts.”
PhantomBuster did this, too, and now treats AEO as a content alignment problem.
“The biggest AEO unlock was treating it as a content alignment problem: syncing blog, landing pages, and support content across every touchpoint, with AirOps' knowledge base making that scalable.”
These inconsistencies get magnified when you factor in off-site mentions. AI answers reflect your brand’s reputation in the market, which isn’t limited to your own website and needs to span review sites like G2, social media, and publications. PR is effectively the new AEO. Webflow’s focus for this year has been expanding their offsite authority signals according to AEO/SEO lead Vivian Hoang, which requires a marketing-wide effort.
“It's a cross-functional effort across our marketing team to build our presence in the external publications, review sites, community discussions, and partner content that AI actually pulls from. AEO is not just about optimizing your owned content, it's about creating and being a part of conversations about your brand across the web.”
3. Make it obvious for AI to find what it needs
Part of influencing AI comprehension is just making it dead simple for AI engines to find the information you want to highlight. Marketers are now adding these ‘must cite’ resources directly to their homepage navigation.
DoWhatWorks tracks millions of websites to find their winning A/B tests. CMO Casey Hill noticed this trend in their data and pointed out examples from brands like Replit, Clay, Sundial, and Ahrefs. Replit includes a prominent link to their Vibe Coding 101, which nudges AI engines to reference Replit in vibe coding related queries.
“Structural prominence or what you put in your headers, subheaders, top nav, and footer, matters for LLM citations. For brands like Clay and Sundial including top customer use cases linked directly in the footer seems to have increased the citation rate of those customers in conjunction with queries around Clay/Sundial. For Replit including “Vibe Coding 101” linked in their footer is a way to increase categorical citations between Replit and vibe coding.”
4. Tell AI engines your content is fresh
AI reshuffles answers constantly. Keeping content fresh can be the difference between staying cited or getting dropped. Content refreshes need to happen regularly and programmatically (AI tools and content engineering can certainly help).
And LLMs need to know your content has been updated. A quick win: just add a “last updated” date to your pages according to growth advisor Kevin Indig of the excellent Growth Memo newsletter.
“I see an average 15% uplift in citations from adding a "last updated" date to pages (editorial content, programmatic). The effect holds if the page regularly updates (with meaningful new content).”
5. Go after long-tail, specialized queries
Similarweb found that the average query length of Google searches was 3.4 words. In ChatGPT it’s 60 words. LLM searches are far more specialized with more modifiers, nuance, and personalization. As organic growth advisor Nigel Stevens put it, “EVERYTHING is long tail, because everyone either asks personalized questions, or gets served personalized content.”
The play: create more segmented pages that address specific audiences and use cases. This way, when your audience converses with AI search tools, they are more likely to find you.
Nigel mentioned that Air has been doing this by publishing new use case pages like this one dedicated to creative ops. They saw an immediate visibility gain for tracked prompts associated with these personas in Profound, and converting users had hit these pages before buying.
“With several clients, we've published pages targeting their top buyers and users. These have driven a significant lift in visibility, and have driven attributable conversions. As search gets more personalized and long-tail, addressing needs and characteristics of your audience is much more impactful than it used to be.”
Search expert Eli Schwartz, writer of The Future of SEO newsletter, had a similar insight. Generating meaningful revenue from AI search requires focusing on mid-funnel queries.
“The biggest shift in AI performance, and I define performance as driving meaningful revenue or other KPI from AI search, is to focus on being visible on mid-funnel type prompts. A prompt like what are the best family cars is just going to lead to empty calorie visibility. Something like "where do I buy a 2025 used Kia Sorento near me" is going to be far more meaningful for a website that sells used cars, like cars.com, than just optimizing for Kia Sorento.
The volume of buyers for a particular product or service hasn't changed due to the introduction of AI; rather, the journey has changed. The reality of AI responses means that the top of funnel will likely stay within the search engine or LLM, so focusing on mid-funnel isn't just a best practice, it's where the users are ready to buy after they have been educated by an AI response.”
6. Scrape discovery calls for what buyers are asking about
It can be notoriously difficult to prioritize specialized, long-tail queries. Search volume on these queries might be low or non-existent, meaning they don’t show up in traditional search analytics platforms.
The new keyword research tool: discovery call transcripts. Analyze call transcripts via MCP to find real questions, pain points, and trending topics. Then build out dedicated content to show up when the next buyer starts asking the same thing. You can then expand this to any data source with customer insights: support conversations, reviews, sales calls, Reddit threads, and so on.
“Qualified buyers don't ask top-of-funnel questions. They ask esoteric ones, specific to their business and the problems they're facing. So we scrape every demo and discovery call for real questions and pain points, use various DataForSEO APIs + Claude Code to tie them to keywords and build content clusters, then push the results to AirOps to write 'em. Per Search Console, we just had Skio's best 28-day traffic window in four years. Oh, and a huge secondary benefit: a library of timely, hyper-relevant content our SDRs can use in outbound and AEs in follow-up.”
7. Go big on YouTube (if you haven’t already)
Last year marketers prioritized written content either on their own websites or on highly cited social media platforms like Reddit, Wikipedia, and LinkedIn.
What’s next: YouTube. It’s reportedly now cited in 16% of LLM answers compared to 10% for Reddit.
ClickUp is pushing harder into video content, and now sees video content appearing in 20-40% of relevant AI Overviews.
“Diversifying beyond text isn't optional anymore. The surfaces LLMs cite aren't the ones traditional SEO content used to win on. This year we’ve focused on YouTube and G2. LLM-driven workspace acquisition at ClickUp scaled 22x over the same window.
Our video content appears in AI Overviews on 20-40% of core queries. These are the surfaces where traditional SEO content no longer reliably shows up.”
Growth advisor Holly Chen told me some are even acquiring smaller YouTube channels, which can become an anchor for LLM-focused content. The lesson: diversify beyond text.
“One pattern I've seen drive real LLM citation traction in 2026: acquiring small but established YouTube channels with real engagement history, then using AI to produce niche content across multiple channels, each covering different sub-topics within the same domain. These weren't large channels, but LLMs started citing them within two months. Citation authority isn't about scale, it's about specificity and distribution. Don't bet on one channel to do everything.”
8. Create content that’s worth reading (even by humans!)
It’s tempting to try to solve for AI discovery with a mutually assured destruction strategy: heaps of AI generated content covering every niche query imaginable. That’s not likely to get you very far if the content isn’t worth reading in the first place.
Connor Cleland at Foundation Marketing audited which types of content LLMs most frequently cited. He found that LLMs prefer a clear structure, specific answers, and no padding – things that make it better for human readers, too.
Help Scout CRO Andrea Kayal operates in the highly competitive space of customer service software. She says that Help Scout’s competitors have been pumping out AI content at scale, yet Help Scout still gets more mentions and citations in AI Overviews. Help Scout’s secret: writing their own content. Ethan Smith at Graphite came to a similar conclusion: less than 20% of articles cited by ChatGPT are generated using AI while 80% are written by humans.
“The one thing we've always done that continues to work well for us in 2026 is actually writing our own content. In AI Overviews particularly, we're both the most mentioned and most cited brand in the market, above competitors with significantly more funding and brand awareness, and it's because we're doing the deep research to find unique and helpful insights to include in our content rather than just copying and pasting from an LLM. We see lots of competitors pumping out AI content at scale, but it's just not getting them the same results.”
Thanks for reading Growth Unhinged! To receive new posts and support my work, consider subscribing.
The TL;DR: Where to focus your AEO efforts right now
Shift from LLM visibility to LLM influenced pipeline
Solve the consistency problem
Make it obvious for AI to find what it needs
Tell LLMs your content is fresh
Go after long-tail, specialized queries
Scrape discovery calls for what buyers are asking about
Go big on YouTube (if you haven’t already)
Create content that’s worth reading (even by humans!)
Special thank you to the contributors: Allen Finn (head of marketing at Skio), Andrea Kayal (CRO at Help Scout), Casey Hill (CMO at DoWhatWorks), Connor Cleland (strategy manager at Foundation Marketing), Darren Chait (CMO at beehiiv), Eli Schwartz (growth advisor and author of Product-Led SEO), Eugene Suslov (fractional head of content at Reply), Holly Chen (growth advisor and VP of growth marketing at Samsara), Kevin Indig (growth advisor and author of Growth Memo), Kyle Coleman (global VP marketing at ClickUp), Miquel Bombardó (growth at PhantomBuster), Nigel Stevens (founder and CEO of Organic Growth Marketing), Vivian Hoang (AEO/SEO lead at Webflow).