Sign in Subscribe

Your Brand Valuation Prices a Customer Who Is Leaving the Room

Your Brand Valuation Prices a Customer Who Is Leaving the Room
Brand equity was built to win a human mind. The AI assistant does not have one to win.

Every major brand valuation method rests on one assumption: that a person makes the final choice. That assumption is about to become the most expensive one in marketing.

Every year the world's most valuable brands are priced to the dollar. Those numbers influence acquisitions, licensing, tax disputes and investor expectations. They are marketing's closest equivalent to a balance sheet.

And every one of them rests on an assumption so obvious that nobody writes it down: that at the moment of purchase, a person chooses.

That person remembers the advertising. That person feels the trust a brand has spent decades building. That person will pay a little more for the name they know. Hand that moment of choice to an AI assistant or shopping agent acting on the customer's behalf, one that holds no loyalty, is not persuaded by advertising the way people are, and weighs evidence rather than familiarity, and the logic underneath the valuation stops holding.

That substitution has started. The question is no longer whether brand valuation is affected. It is how long the industry can keep valuing brands as though it isn't.

Every valuation contains a hidden customer

The established methods differ in mechanics but share a foundation. Interbrand estimates the role a brand plays in the purchase decision and multiplies it by the brand's strength and earnings. Brand Finance asks what a company would pay to licence its own name, and adjusts that royalty by a brand strength index built partly on consumer familiarity, consideration and preference. Kantar BrandZ asks consumers whether a brand is meaningful, different and salient, and links the answers to financial performance. Behind much of modern marketing science sits the Ehrenberg-Bass idea of mental availability: the brand that comes to mind in a buying situation is the brand that gets bought.

These are serious, tested frameworks, and they have earned their place in boardrooms and courtrooms. But look at what each one is actually measuring. Role of brand in the purchase decision. Familiarity. Preference. Salience. What comes to mind. Every input describes the inside of a human head at the moment of choice.

When the decision is made by an AI assistant or shopping agent, what exactly is the role of brand?

That is not a rhetorical question. It is a measurement problem, and none of the existing instruments is built to answer it.

The new customer is not persuaded the way people are

In September, AIVO ran 2,160 buying conversations in which a small business asked ChatGPT, Gemini and Perplexity to help it choose a bank. Fintechs received 1,345 recommendations. Traditional banks received 452. On ChatGPT and Gemini, a single fintech, Mercury, was recommended 383 times, more than every traditional bank in the study combined.

Think about what those banks bring to that conversation. Brands built over a century or more. Branch networks, sponsorship, advertising budgets measured in hundreds of millions, and in several cases a place among the most valuable brands in the world. By every traditional measure, these are some of the strongest brands in commerce. In the conversation where the customer actually asked for a recommendation, a company founded less than a decade ago beat all of them put together.

This is not a banking anomaly. Across more than 20,000 multi-turn buying conversations covering over 200 brands, AIVO has found that 87.3% of brands present early in a conversation are displaced before the final recommendation. Being known to the model is common. Being chosen by it is not.

Brand equity was built to win a human mind. The AI assistant does not have one to win.

The assistant is not hostile to famous brands. It simply weighs different things: specific evidence, recent reviews, product pages, third-party comparisons, fit to the stated need. Some of what made a brand strong carries over. Much of it, the warmth, the familiarity, the habit, does not.

The problem is in the forecast, not the present

The obvious objection is scale. AI assistants still account for a small share of buying decisions. In financial services, AI referrals make up well under 1% of website traffic. Nobody has yet measured how strongly an AI recommendation changes what people finally buy. On today's numbers, the effect on any single brand's revenue is modest.

That objection is correct about today, and it misses the point. A brand valuation is not a measure of today. It is a forecast: years of projected brand earnings, discounted back, plus a terminal value that assumes those earnings continue. The present is a small part of the number. The future is most of it.

And the future is where the shift is happening. EY's 2026 survey of more than 18,000 consumers found that 14% had already let AI select a financial provider on their behalf. G2's survey of B2B software buyers found that 69% had chosen a different vendor than planned because of an AI chatbot. Whether those numbers double in three years or in ten, they sit squarely inside the forecast period of every brand valuation published this year.

A finance director would never accept a discounted cash-flow model that silently assumed interest rates, tax or inflation would stay unchanged for a decade. Yet today's brand valuations silently assume the decision-maker remains human throughout the forecast period. That assumption deserves the same scrutiny.

A valuation that assumes the human customer persists for the next decade is making a forecast. It is just not disclosing it.

No one is suggesting the established tables are wrong. The suggestion is narrower and harder to dismiss: they contain an undisclosed assumption about who makes the choice, and that assumption is now testable.

The missing term

The answer is not to throw out the established methods. They value something real: what a brand is worth in human minds, and the legal and financial frameworks built around that value will not disappear. The answer is to add the term they are missing. In finance, every major shift eventually becomes another factor in the model. Brand Alpha is simply an attempt to measure one that did not previously exist.

The missing term answers a single question: when an AI assistant or agent makes the recommendation, does this brand perform above or below the position it holds in the market? A brand recommended twice as often as its market share predicts has an AI advantage. One recommended only a fraction as often has an AI liability. The economic value of that difference, measured across real buying conversations, is what AIVO calls Brand Alpha.

It is deliberately built to sit beside the existing tables, not to compete with them. It uses market share as its baseline, so it inherits everything the established methods know about a brand's position. It measures only what they cannot see: what happens to that position when the customer hands the decision to a machine.

The methodology is public, and we would rather it was challenged than ignored. Brand valuers, finance teams and academics are invited to test it.

The cost of not looking

For forty years the question in brand valuation has been how much a brand is worth to the people who buy it. That question is not going away. But a second one has arrived beside it, and the brands that ignore it will find out the answer the expensive way.

The question is no longer only what your customers think of you. It is what their assistant thinks of you, and whether anyone in your boardroom has measured it.

Sources

AIVO Meridian

Comments ()