Most advice about AI search visibility is written as though the major assistants behave the same way. They do not. The composition of their source pools differs sharply, and strategies that produce results on one platform can produce almost nothing on another.
Royston G King argues that treating the assistants as interchangeable is the most expensive error in the category. A Forbes 30 Under 30 Monaco honouree and University of Southern California alumnus accepted into Columbia University, who bought his first stock at fourteen and started his first company at seventeen, he founded Master Scaling in 2018 and Quantum Scaling Partners as its selective arm. He has watched organisations spend against a single playbook and see returns on one platform only.
The clearest illustration comes from two products built by the same company. Analysis of citation patterns has found Reddit accounts for a substantial share of social citations in Google’s AI Overviews while representing a far smaller fraction of what Gemini cites. Two systems, one organisation, materially different behaviour. Any strategy assuming uniformity is already misallocating budget.
A few patterns are worth understanding before committing spend.
Perplexity leans heavily on community content and freshness. Independent tracking has consistently found community platforms making up the largest single-domain concentration of its citations, alongside an unusually high share of academic and institutional sources. It also produces more citations per response than most competitors, which means more available slots. Recently published threads can surface quickly, making timeliness more valuable here than anywhere else. The caveat is platform dependency risk. When one major community platform entered litigation with Perplexity, its citation share dropped sharply and video content partially filled the gap. Channels that depend on a single platform relationship carry structural fragility.
Google AI Overviews weight community and video content heavily. Video has climbed steadily in tracking studies, and a meaningful share of AI Overview citations come from video URLs that do not rank in the top 100 organic results for the same query. That is a significant opening. It means a well-made video with an accurate transcript can earn citations without competing in conventional rankings at all.
ChatGPT draws on community platforms, encyclopedic references, and business press. A large portion of its most-cited pages sit on domains no brand can pitch its way into, including reference works, government sites, and academic institutions. Practical strategy competes for the remaining share rather than attempting the impossible.
Notably, a substantial fraction of pages ChatGPT cites have little or no conventional search visibility, which means traditional rank tracking will systematically miss them.
Gemini behaves closer to an encyclopedic model. Reference-style sources and Google-owned properties carry more weight, and community content carries considerably less. Effort that produces results on Perplexity may barely register here.
Three operational conclusions follow.
First, decide which engines matter for the specific business before building a plan. A company selling to consumers who research on their phones has a different priority list than a firm selling to procurement teams. Spreading effort evenly across all four platforms guarantees mediocre performance on each.
Second, stop treating conventional search rankings as a proxy. Studies examining thousands of prompts have found only a small minority ofAI-cited URLs appearing in the top ten conventional results for the same query. A company can rank well and remain invisible in AI answers, or rank poorly and be cited frequently. These are different systems measuring different things.
Third, measure distributions rather than single results. The same query can return different sources on repeated runs, with variation as high as half the time across different regions. One check of one prompt is noise. Useful measurement means running a defined set of prompts repeatedly and tracking how often a brand appears across the distribution.
The underlying point is that AI visibility is not one market. It is several markets with different rules that happen to share an interface metaphor. Treating them as interchangeable is the most common and most expensive error in the category right now.
AI visibility, as King puts it, is not one market but several with different rules that happen to share an interface. Quantum Scaling Partners builds engine-specific plans for that reason rather than a single strategy applied evenly.
To learn more about Royston G. King, visit his official website. You can also follow him on Instagram, connect with him on LinkedIn, and watch his content on YouTube.
