Introduction
Most brand-growth theories are built around attention, memory and availability. However, these don’t answer why some brand memories become valuable. Meaning, culture, identity, emotion, trust, social proof and experience drive the content and quality of those memories. In other words, availability puts the brand in the choice set; meaning and experience determine whether it should be chosen.
Brands have invested in reach, unique assets, emotional consistency and salience because these have made them more memorable and easier to choose. AI doesn’t upend that logic but layers a filter on top of it. Consumers are delegating more decisions to AI to search, shortlist, compare, summarize and recommend. In the future, consumers may not be involved in those decisions at all. A world of AI agents deciding for us is entirely possible. Brand visibility now depends not just on whether consumers can remember your brand, but whether machines can recognize it, retrieve trusted information about it, compare it accurately, and explain why they should recommend it. Rather than striving solely for mental availability, brands must now develop mental, physical, and AI availability. This leads us to the brand challenge of the next decade: how do we build equity and emotional resonance with people, while also becoming legible and trustworthy to machines?
The theory stack
We think the strongest way to frame this is not to treat AI visibility as an entirely new marketing problem. It is better understood as an extension of established brand-growth theory. Here is a brief review of some of the relevant and pertinent theories as a starting point.
Binet and Field: long-term brand effects
Binet and Field’s IPA work shows that brands need to balance short-term activation with long-term brand-building. Their work focuses on how campaign effects develop over time and warns against relying too heavily on short-term online metrics at the expense of long-term brand strength.
Branding builds memory structures, fame, emotional connections, etc. Activation flips existing demand, but without brand-building it rarely produces long-term growth.
In this discussion, we see this as applicable to AI visibility as well. When AI systems are trained on and retrieving information from the environment around brands, those brands with a larger cumulative presence are probabilistically more likely to be mentioned in AI-generated responses.
Byron Sharp and Ehrenberg-Bass: availability
Byron Sharp’s contribution is different. His core argument is not primarily about the brand/activation budget split. It is primarily about availability. The Sharp and Ehrenberg-Bass theory argues that brands grow by increasing penetration, which requires both mental availability and physical availability. Ehrenberg-Bass also links mental availability to Category Entry Points: the needs, occasions, motives and contexts that trigger category consideration. A brand grows when it is linked to more of these buying situations.
Charlie Oscar/WARC: AI visibility
Charlie Oscar/WARC gives us a new reflection on the AI-era evidence layer. The work attributes two-thirds of LLM visibility to long-term brand equity, with current marketing activity and citations making up smaller shares. WARC’s broader guidance also argues that brand fundamentals become more important, not less important, as AI tools insert themselves into the relationship between consumers and brands. Interestingly, Charlie Oscar/WARC suggests that AI systems may now reward the same cumulative brand effects by making established, well-evidenced brands more visible in AI-mediated discovery.
A brief comparison between Binet and Charlie Oscar’s work highlights some very interesting parallels.
As mentioned, Charlie Oscar’s work shows that approximately two-thirds (≈66%) of brand visibility inside Large Language Models (LLMs) is driven by long-term brand equity. This closely mirrors the long-established IPA / Les Binet finding that around 60% of marketing effectiveness comes from long-term brand building. This is not a coincidence. LLMs structurally reward the same cumulative, long-term effects that econometric models observe in brand growth and profitability.
LLMs do not rank pages or optimize for bids. They synthesize answers based on:
As a result, AI visibility appears to reflect a combination of brand equity, source consistency, and the quality of machine-readable documentation across the information environment. Scale helps, but structured evidence and credible third-party corroboration can alter the outcome. This means smaller brands may still achieve disproportionate AI visibility if they are consistently documented across authoritative and machine-readable sources.
There is a direct structural equivalence between Binet’s work on brand effects and LLM behavior:
| Binet Effect | LLM Analogue |
|---|---|
| Long-term brand equity | Historical data accumulation |
| Mental availability | Generation probability |
| Broad reach | Cross-context mentions |
| Distinctiveness | Stable entity representation |
| Consistency | Reinforced linguistic signals |
| Scale | Probabilistic dominance |
Binet’s 60/40 rule refers to marketing effectiveness and profit growth. LLM visibility percentages refer to how discovery systems allocate attention. LLMs do not suggest brand now matters more. Rather, they surface brand effects earlier in the customer journey, before activation begins. AI does not replace brand-building theory. It adds a new visibility layer that Marketing and Research leaders must measure. AI visibility appears to behave like a new form of availability. Brands must now be mentally available to people, physically available in channels, and machine-available to AI systems.
AI availability: a new concept
Brand visibility in AI-mediated environments increasingly depends both on broad machine legibility and on how personal agents learn, retain and reproduce individual consumer preferences over time. As brands strive to become machine-legible through websites, product data, reviews and FAQs, AI availability can be defined as the probability that an AI system can correctly identify, retrieve, understand, compare and recommend a brand in response to a relevant consumer prompt.
It is not the same as awareness. A brand can be famous among people but poorly represented in AI answers if its claims are unclear, its content is hard to parse, its proof points are weak, or its third-party corroboration is thin. It is also different from SEO. Traditional SEO helps a brand become discoverable in ranked search results. AI availability is broader. It includes whether the brand is present in the sources that AI systems trust, whether its information is structured, whether it is consistently described across the open web, and whether it can be confidently associated with particular category entry points. AI tools are not only drawing from historical training data, but many also now retrieve live information from the web. This means brands are not only competing inside model memory, but they are also competing inside the live information environment.
Why big brands will not always win AI search
The obvious assumption is that large brands tend to have more equity, therefore LLMs will surface large brands more often. That is directionally true, but it is incomplete. Large brands have several advantages. They tend to have more historical mentions, reviews, media coverage, backlinks, consumer discussion, marketplace presence and comparison content. These signals may increase both model familiarity and retrieval probability. If large brands receive more attention from LLMs, is that because of brand equity itself, or because larger companies have more resources to optimize their information environment? The answer is likely both.
Large brands may be more visible because their brand equity produces more cultural and digital traces. However, AI visibility can also be affected by tactical factors: FAQs, structured pages, product data, review environments, third-party articles, category guides, comparison content and paid placements.
M+C Saatchi Performance describes this as a shift where traditional SEO is no longer enough. Their analysis argues that brands can perform well in search but remain absent from AI answers if they lack corroboration across directories, review platforms, roundups and trusted third-party sources.
AI does not simply reward size. It rewards brands that are salient, sourceable, consistent and easy to justify. Scale helps, but structured evidence and credible third-party corroboration can alter the outcome.
Category Entry Points become prompt entry points
We know that in Sharp/Ehrenberg-Bass thinking, brands build mental availability by linking themselves to Category Entry Points. In an AI world, Category Entry Points increasingly become prompt entry points. These prompts are modern expressions of category entry points. The AI system then acts as an intermediary, translating the user’s need into a shortlist, explanation or recommendation.
This changes the research problem. Brands should not only ask whether they are known, they should ask:
For which prompts does the brand appear?
For which category needs is the brand associated?
What language does the AI use to describe the brand?
Which competitors are mentioned alongside it?
Which sources are shaping that answer?
Does the AI frame the brand how we want it to be framed?
This is the AI-era version of brand salience research. The unit of analysis moves from the survey association to the prompt-response environment.
Conclusion
AI visibility should not be understood as replacing established brand-growth theory, but as extending it into AI-mediated discovery environments. Mental availability, physical availability and AI availability may increasingly operate together as interconnected layers of modern brand visibility. The next challenge for marketing and insight teams is understanding how brands become retrievable, interpretable and recommendable inside generative systems.
Want to learn more about marketing in an AI world? This article is the first in our three-part series. In the second blog, we explore how brands can drive AI Availability through influencing decision agents and LLMs. The final installment examines the practical and strategic implications for building a machine-readable brand.
For a deeper exploration of this topic, stay tuned for our upcoming whitepaper.
Author: Jon Arthurs
Managing Director, Eastern Europe at Toluna
This series is written by Jon Arthurs, Managing Director, Eastern Europe at Toluna and with thanks to Rick Candelari, Global Team Lead Solution Consulting at Toluna for additional contributions to these articles.
