AI platforms have changed the way people search, and winning on these platforms requires a strategic rethink. Our research shows that when someone types a query into ChatGPT or Gemini today, both engines are asking 10-20 related questions behind the scenes.
That single keyword explodes into vendor comparisons, technical implementation questions, platform evaluations, and trial queries. Companies basing B2B content marketing strategies around single keywords and traditional methodology are losing huge amounts of potential traffic.
Read on to discover how AI search really works, how to build a strategy for both AI and trad search, and the data needed for measurable results.
How personalized search leads to query fan-out
The way people search has fundamentally changed. With legacy search engines like Google, users learned to keep queries short and generic – typing things like ‘HR platform’ because that’s what worked best with keyword-based algorithms. Google would then try to guess their actual intent, serving up articles that attempted to cover multiple possible meanings. This led to those bloated ‘ultimate guides’ we all know too well – content trying to be everything to everyone.
AI platforms have flipped this completely. Users can now ask specific, detailed questions because they know they’ll get tailored answers. Instead of ‘HR platform’, they’re asking:
- ‘What HR platform handles remote teams across multiple time zones?’
- ‘HR platform that integrates with Slack and doesn’t require IT setup’
- ‘Compare BambooHR vs Gusto for 50-person startup’
- ‘HR platform GDPR compliant with EU payroll’
- ‘Free trial HR software no credit card required’
This shift in user behavior – from generic keywords to specific, conversational queries – demands a new content strategy. We call this approach “query fan-out.”
Query fan-out is about recognizing that a single head topic now branches into dozens of specific queries that people are actually typing. Rather than creating one comprehensive article trying to address all possible intents, we create highly targeted content that directly answers each of these detailed queries.
- The old approach: One keyword = one article covering multiple intents.
- Our approach: One head topic = multiple pieces of content, each addressing a specific query variation.
This strategy works because AI platforms are more likely to surface and cite content that precisely matches what users are asking. When someone asks “HR platform that integrates with Slack and doesn’t require IT setup,” they’re more likely to get an answer sourced from content that specifically addresses that exact need – not from a generic “Complete Guide to HR Platforms.”
Our research has uncovered how different searcher profiles approach the same general topic with vastly different specific queries. Take ‘migrate to cloud’ as an example – this single topic generates distinct queries from at least four different user types:
- Research intent: ‘benefits of cloud migration for small businesses’
- Comparative intent: ‘AWS vs Azure migration tools comparison 2025’
- Emergency intent: ‘cloud migration failed rollback procedures’
- Planning intent: ‘cloud migration timeline template 50 employees’
Each of these queries deserves its own focused piece of content, not a section buried in a massive guide.
If you’re still optimizing for single keywords and writing broad articles that try to be everything to everyone, you’re missing the real searches happening on AI platforms. The opportunity lies in understanding all the related specific queries under your head topics and creating dedicated content for each one.
AI search platforms uncover value
We discovered something fascinating while analyzing thousands of AI responses – educational and technical content gets cited far more frequently than marketing pages.
AI systems crave depth and specificity, meaning glossy landing pages with marketing-led content are often ignored. Whereas dense technical documentation powers dozens of ChatGPT responses daily. Ultra-specific niche content gets cited by AI at rates that defy traditional SEO logic.
Take our client Divi Life. When users search ‘divi mobile menu full screen’, over content, which covers how to do complicated things in Divi, gets pulled directly into AI Overviews with step-by-step instructions.
As of August 2025, 38% of citations across Divi Life’s most relevant topics come from Ellipsis-produced content. Content quality is key here. We write with the user’s needs in mind, answering specific questions directly, and in a language that matches their know-how. Our Divi Life articles get cited because they have the depth that both users and AI platforms are looking for – and we wrote the majority of these before we even started optimizing for AIOs and LLMs.
“We’re seeing clients gain 100+ new AI citations monthly while their competitors remain completely blind to what’s happening. They don’t even know they’re losing because they can’t measure it.”
– James Baldacchino, Head of Strategy & SEO, Ellipsis
This doesn’t mean that technical documentation is the key to winning at AI search. It just means that AI platforms have the nous to identify ultra-relevant searches. They can see the value in that useful product-specific info you have on your site that Google could never decide what to do with.
When someone asks ChatGPT a complex B2B question, they want specifics, not marketing fluff. AI systems know the difference, so this needs to be factored into any content strategy.
Building content strategies for AI search
The good news is, AI search is an end to the days of regurgitated, ‘goog’-enough content that still dominates the SERPs. For some time, users have been scrolling through ‘cover everything’ content in an attempt to find the one or two sections that answer their query. Too many SERPs are a mausoleum of mediocrity, and AI platforms have opened the door to genuinely useful content.
In building strategies for our clients, we focus on topics over keywords, relevancy over search volume, and use proprietary software to track the highest priority queries matched to our clients’ ICPs. We then monitor our client’s performance for these queries over time.
Think beyond keywords
As search intelligence has improved, keywords are less relevant, with topics and queries taking over. Your ideal customer cares more about answers, and they don’t care whether they find them through Google or by asking ChatGPT at 2am.
Where search volume fits in
Search volume is becoming less important when taken in isolation. A 50-monthly-search term might be exactly what your highest-value prospects need.
We’re finding massive opportunities in ultra-niche, technical queries that traditional SEO would dismiss. Think about queries like ‘Kubernetes RBAC misconfiguration detection tools’ – minimal search volume, but the people searching are DevOps leaders with serious budgets. These ultra-specific searches appear in far more AI conversations than their numbers suggest.
Search volume offers an excellent starting point, but words and phrases need to be looked at in a much wider, user-led context.
Here’s how we determine if a keyword is a good fit for our clients. Before diving into any analysis, we first need to deeply understand who we’re targeting. We go through the client website, products, and competitors to build a comprehensive understanding of their ICP. We use our dashboard to generate detailed ICP and company profiles, ensuring we know exactly who the ideal customer is.
From there, we identify the most important keywords – these are core topics central to the client’s business, keywords they’re already ranking for in Search Console, and anything crucial to their products and ideal customers.
We then find keywords from competitor gaps (keywords that competitors have content for but our client doesn’t) that are good ICP matches. We also look for missing opportunities – topics that are perfect ICP matches but haven’t been covered by the client or their competitors yet.
Once we have this comprehensive list, we move forward to the analysis process using FALCON AI, our proprietary software now in its third generation.
With our prioritized keywords in hand, we take a broader view of the search world. We fetch the full SERP, scrape the top organic results, pull every citation from AI Overviews, and extract all People Also Ask content.
This data is then run through our FALCON-trained models, along with insights about our client’s business and ICP, producing 10-20 natural language variations for each keyword. These suggestions are shaped to reflect real user behaviour and filtered specifically for the ideal customer, leaving us with the actual queries prospects are asking.
FALCON has been completely reengineered to optimize for the new SEO reality. By analyzing thousands of queries and the content behind them, we’ve identified clear patterns about what Google surfaces, what AI Overviews cite, and what types of content consistently win visibility. These insights shape every content strategy we create.
The result is a content strategy that works with the fan-out effect of AI platforms, positioning our clients to appear in search results and the AI-powered answers that users increasingly rely on.
Traditional search would have you compete for ‘CRM software’ against Salesforce’s massive monthly content budget. Whereas AI-search has opened the doors to brands being able to identify dozens of ultra-specific scenarios that their exact buyers search for.
Reframing success
Most companies have zero visibility into their AI performance. We’re tracking citations and mentions separately and benchmarking against competitors weekly.
OpenAI recently made model selection invisible to users, running most queries on GPT-5 Mini for efficiency, over the flagship GPT-5. Content that works for one model might fail on another, so we made sure our tracking monitors performance across model variations, something literally nobody else is doing right now.
Here’s what all this tracking reveals:
- Mentions are worth more than citations (brand awareness vs. authority).
- Certain prompts drive significantly more value despite similar search volumes.
- You can dominate AI while losing in traditional search (and vice versa).
- Competitor content from years ago might be beating your current content in AI responses.
Google, OpenAI, and Anthropic aren’t releasing this data, so any tool claiming to have exact AI search volumes is guessing. That’s why we built our own methodology. It’s the only way to get real data about where your content appears across ChatGPT and AI Overviews.
We’re achieving 24-25% AI presence gains in 29 days for our clients, while their competitors are still trying to work out how to track and report on this.
By creating our own in-house solution, we’re able to give our clients a granular level of performance tracking across AI platforms, as well as search engines. Without our technology, clients simply cannot access this data, leaving them completely in the dark regarding their overall search performance.
‘SEO Alligators’: Clicks vs impressions
A client of ours recently gained 24% impressions month-on-month while flatlining on traffic. Impression explosions are a common pattern across our client accounts over the past 12 months:
These impressions directly correlate with AIO citations. Here’s an example of what we’re seeing from our clients, with traffic and impressions crossed over to form a hungry gator’s jaws:
Technical docs or product feature studies that barely get clicked might be getting cited across ChatGPT and Perplexity constantly, featured in AI Overview responses for high-intent queries, and driving enterprise deals where prospects say ‘ChatGPT recommended you’.
Our tracking methodology prioritizes opportunities through the following:
- ICP match score (how well does this query match your ideal customer?)
- Funnel position (bottom-funnel queries get higher priority)
- Competitive gap (where can you dominate vs. competitors?)
We calculate this for every single query, then group them to find topics with the most potential. Our final list shows exactly where you can win fast for topics that actually convert.
Smart distribution in the AI era
While content remains powerful for SEO, smart companies multiply their impact through repurposing. A comprehensive guide you’ve published can be transformed into:
- 6 LinkedIn posts highlighting key insights.
- 3 newsletter segments diving deeper into specific points.
- Video scripts for YouTube or webinars.
- 10 Twitter threads for bite-sized consumption.
- Slide decks for sales enablement.
Look to embrace both ‘push’ and ‘pull’ distribution channels to spread your reach.
Push distribution actively delivers content to the places where your audience already spends time, like on social media or through newsletters. Pull distribution ‘pulls’ users to your brand through online searches, and comes from publishing content that directly answers what your audience wants to know.
Explore channels that are relevant to your audience. Pick channels based on where your actual buyers spend time, not where marketing blogs tell you to be.
Futureproof your B2B content strategy with Ellipsis
Content marketing success now demands excellence across Google, AI platforms, and countless distribution channels. Our Content Growth service delivers:
- Topic calendars that capture query fan-out: Each topic maps to 10-20 AI queries.
- Fan out keyword methodology: We identify every intent variation to dominate topics.
- Weekly AI performance tracking: 500-700 prompts monitored per client.
- Competitive benchmarking: See exactly where competitors are winning, and why.
- Speed to market: 100+ new citations monthly.
Our approach rests on three pillars:
- Process: Data-driven quality at scale.
- FALCON AI: The only platform that truly tracks AI search performance.
- Expertise: In-house writers and strategists who understand search and B2B buying.
Winning at AI search is about creating solutions and taking measurements over guesswork. Our clients are already locking in positions that will define their market presence for years to come, and there’s no time to waste in getting started.If you’re ready to build your own content strategy, book an AI SEO Audit today.
