Waikay’s Query Fan Out Feature
What it is and why it’s essential for your AI optimization toolkit
The Basics
Tackling AI optimization is tough. It’s new, it’s confusing, and there’s no established playbook yet. Most of what we know comes from early experiments and people sharing what’s working for them in real time. On top of that, a wave of tools is trying to frame AI optimization as though it’s the same as SEO keyword research—selling “AI search volumes” and treating them like the old search metrics we’re familiar with.
But here’s the thing: AI doesn’t generate search volumes. There’s no database of how many people asked ChatGPT or Gemini the same thing last month. Even if you run the exact same phrase through AI multiple times, the responses can vary significantly. That’s just how it works—it’s generative, not indexed.
So if we can’t rely on volume numbers, how do we actually improve our visibility? The only sustainable approach is to study patterns over time across a broad range of related queries. Not just hammering on a single keyword. Think about it: if you want to rank for “dog food,” saying “dog food” a million times on your site won’t work. (It hasn’t in SEO for a long time.) Yet somehow, as the conversation shifts to AI search, people seem to have forgotten this basic principle.
To understand what “ranks” in AI, we need to step back and look at the foundations of how AI works. And that foundation is entities.
Entities in AI
Entities aren’t just words; they’re concepts. They encapsulate multiple dimensions of meaning. Take “dog food” for example. As an entity, it immediately pulls in related ideas like health, ingredients, fish, chicken, broth, price, quality, pets, even emotional associations like “happy dog” or “cheap but unhealthy.” Humans don’t process “dog food” as a standalone word—we automatically think of the context around it. AI does the same thing.
That’s why extracting insights from AI is less about exact numbers and more about qualitative patterns. It’s almost like talking to people. If you asked 100 people on the street, “What’s the best dog food?” you wouldn’t get one single answer—you’d get 100 slightly different ones. Maybe one person swears by a cheap supermarket brand, another swears by a premium organic option, and a third admits their dog loves the unhealthy stuff but they only buy it occasionally.
The point is: those answers, in aggregate, give you the real picture. You’d be able to see the frontrunners, the themes, the trade-offs people make. And this is exactly how we need to treat AI.
What is Query Fan out?
So, back to the big question: what does this have to do with Query Fan Out? And what does that even mean?
Query Fan Out is about refusing to stay stuck on just one keyword or phrase. Instead, it deliberately expands—fans out—into the related entities and contexts that humans naturally introduce when they talk about a subject.
Take the dog food example again. A single person might answer:
- “Well, price-wise Brand X is best, but quality-wise I prefer Brand Y.”
- “My dog loves this one, but it’s full of fillers, so I only buy it as a treat.”
- “I used to buy that brand because it was cheap, but my vet told me it wasn’t healthy, so I switched.”
Notice how instantly they’re introducing qualifiers—price, quality, health, convenience. Those aren’t just throwaway details; they’re the entity markers that define how people (and AI) evaluate the space.
Query Fan Out anticipates these kinds of qualifiers. Instead of you having to manually think of every possible angle, Waikay uses data on entity relationships to automatically expand your seed query. That way, you can see where your brand sits across the multiple dimensions AI systems are likely to consider.
Waikay’s Query Fan Out explained
So, how does Waikay actually put this into practice? Thankfully, it’s more straightforward than it sounds.
- You start with a phrase you want to rank for—for example: “What is the best dog food brand?”
- Waikay queries an AI system (like Gemini) and pulls back variations of that phrase—different ways users might frame the same question.
- Those variations often introduce new angles: health, affordability, nutrition, convenience, sustainability, breed-specific needs, etc.
- You then select the ones that make the most sense for your brand and add them to your query pool.
The end result is a full picture of your brand’s presence in AI search. Instead of a shallow snapshot tied to one phrase, you’re building an entity-rich dataset that reflects how AI—and by extension, users—are actually talking about your space. It means you don’t have to know whaty to track, but you have all the tools to discover what to track.
Example:
Here’s an example of the query fan out feature on Waikay. You can see that, alongside attacking angles of different entities, it also starts to span across different audiences. For example, General Rankings are much different to Expert or User reviews. It allows you to start to see your brand from different perspectives, and introduces different things people consider when searching for authority.
That’s the real power of Query Fan Out: shifting from keyword obsession to context awareness, and from single snapshots to holistic understanding.
To Access this, first head to the ‘brand visibility’ feature:
Then create a prompt. Make sure it is a funnel query that allows for multiple brands to show up in the results
This will automatically trigger the query fan out. Use the query fan out to add to the prompts you track. Start from a single idea and grow based on logic, not search volume.
But I want Search Volume!
Now, a lot of SEOs immediately ask: But is this data accurate?
The answer: yes—but not in the old keyword-metric sense. It’s accurate in the same way surveying 100 real people is accurate. AI is trained to mimic human communication, so when we analyze its responses qualitatively, we’re essentially running a massive, automated focus group. It’s messy, it’s human-like, and it’s far more realistic than chasing nonexistent “AI search volume” numbers.
But where do I get This Authority?
So many conflicting sources on this at the minute. Our best understanding Is repeated and consistent feedback from multiple sources. Reddit was (and, accoriding to Waialky) is a huge citation pool for brand authority- but we cannot be but we cannot be solely reliant on Reddit or any single platform. Authority in AI responses—especially for enterprise or commercial use—requires a multi-pronged strategy rooted in transparency, consistency, and verifiable expertise.
• Citations from trusted domains: Responses that link to recognized, high-authority sources (think .gov, .edu, major publications, or industry leaders) carry more weight. Reddit can be useful for sentiment and community validation, but it’s not a primary source of factual authority.
• Schema and structured data: If your brand is publishing content, embedding schema (especially , , and properties) helps search engines—and AI models—understand your credibility. This is especially true for platforms like InLinks that automate internal linking and schema deployment.
• Consistency across platforms: Authority is reinforced when your brand’s messaging, expertise, and tone are consistent across your website, social media, documentation, and third-party mentions. AI models pick up on this repetition.
• Expert-authored content: When your brand publishes content attributed to real experts (with bios, credentials, and external validation), it’s more likely to be cited or referenced by AI systems. Authority is earned through clarity and depth, not just volume.
• Engagement and backlinks: If your content is referenced by other authoritative sites, especially in technical or niche domains, it signals trustworthiness. AI models often weigh backlink profiles when determining which sources to trust.
• Transparency and disclosures: Especially in sensitive domains like health, finance, or security, clear disclosures, update timestamps, and editorial standards boost perceived authority.
Truthfully? Authority isn’t something you “get” once—it’s something you build, reinforce, and protect. AI systems are trained to favor sources that demonstrate expertise, trustworthiness, and relevance over time. That means your brand needs to show up consistently, speak clearly, and back up claims with evidence.
