Key points
- AI is creating an inferred brand beyond the brand we deliberately express.
- The inference gap is the distance between those two identities.
- The INFER framework helps us understand and manage what AI concludes about us.
We all have an individual brand. Our clothes, words, and ideas all project signals about who we are or who we want to be.
I’ve spent more than 25 years building brands, including senior roles at Ogilvy, where much of my work involved a deceptively simple yet critical question: How do you shape what people think about a company, brand, or individual? Of course, what a brand intends to communicate and what people actually perceive have never been exactly the same. That gap is as old as branding itself and has often kept me up at night.
But AI is changing the calculus because the interpreter is changing.
From Expression to Inference
Traditional branding is largely overt, and that’s by design. Companies choose hallmarks that include things like names, logos, colors, typography, imagery, and messages. These are often crafted with extraordinary attention and enormous budgets. The best brands can go deeper and create emotional connections that can make a swoosh or an apple carry meaning far beyond color and typography. It’s the formation of a relationship between the user and the brand.
AI encounters a brand very differently. Large language models can draw upon corporate websites, journalism, reviews, research, social commentary, historical references, and just about any online information. Some of those signals are carefully created by the organization, while many are not. AI gathers this material and constructs something like a probabilistic portrait that can be far from the “brand persona” that a company or person intends.
This creates what I think of as two versions of a brand. There is the expressed brand, the identity we deliberately create and communicate based on traditional brand hallmarks, and increasingly there is the inferred brand. That’s the identity AI reconstructs from the signals, intended or unintended, surrounding us. The interesting question, for me and most marketers, is how closely those two portraits match.
Being Found Isn’t Being Understood
Business has long used frameworks such as SWOT (strengths, weaknesses, opportunities, and threats) and PEST (political, economic, social, and technological) to understand an organization and the world around it. The digital era brought SEO (search engine optimization), and now GEO (generative engine optimization) asks whether AI and LLMs can find and surface a person or organization. These tools answer useful questions, but being found is not the same as being understood. None of them quite capture the identity AI constructs once it has gathered the available signals.
Think about a physician who has spent 20 years building a reputation around a particular clinical area. Her website reflects it, her colleagues know it, and patients may seek her out specifically because of it. Yet ask a few LLMs about her and another portrait might emerge. Perhaps an older affiliation or even a resolved malpractice claim from the past dominates the answer. Her research is barely mentioned, or she is described in the same generic language used for dozens of physicians in the field. And the same applies to large corporations where issues outside the conventions of branding drive AI output.
Nothing necessarily has to be factually wrong for the portrait to be wrong. The problem can simply be the distance between the identity she has spent years building and the identity AI has inferred from the available evidence. Once that inferred version begins circulating first, the expressed self can find itself arguing with a portrait that never truly existed.
That raises a new question for individuals and organizations alike: What does AI infer about you?
A Framework for Inference
Leveraging my brand-building work in the past, I’ve been thinking about that question through a framework I call INFER. It offers five ways to examine the identity that emerges when AI interprets the signals surrounding a person, company, or organization.
- I—Identity. Look at the portrait itself. What does AI think you are?
- N—Narrative. Look across different questions and contexts. What story keeps emerging, and what gets left out?
- F—Fidelity. Compare that portrait with the identity you intended to create. How close are they?
- E—Evidence. Where did this version of you come from? The sources and signals shaping the inference may be very different from the ones you expected.
- R—Resilience. Change the model, change the question, or come back six months later. Is the identity still recognizably you?
Artificial Intelligence Essential Reads
Why ‘Relational Foreclosure’ Matters in the Age of AI
The Educational Artifact Is No Longer Enough
INFER isn’t intended to replace SWOT, PEST, or the traditional tools of branding. It examines something those frameworks were never designed to see, and that’s the distance between the identity we express and the identity machines infer. That’s the inference gap, and it’s critical to a business or an individual.
When AI Becomes the Interpreter
AI is a curious if not quirky kind of audience. It doesn’t buy your product or walk into your store. It interprets you for people who might. Someone can ask an LLM which company is most innovative in an industry, which university has the strongest academics, which physician is known for a procedure, or which car mechanic is better. Increasingly, the response arrives as a coherent answer rather than a page of links that requires the person to do the interpreting.
This reaches well beyond corporations. Universities, hospitals, nonprofits, executives, scientists, physicians, authors, and perhaps all of us increasingly have identities that can be reconstructed by AI. The accumulated record of “us” becomes part of the informational corpus from which those telling inferences are made.
For years, we have thought carefully about the brands we present to the world, from the colors and symbols we choose to the messaging we hope will define us. AI adds another layer because our identity is increasingly being encountered by machines that gather those signals and construct a version of us that others may encounter first.
Branding has always been about crafting a meaning, real or perceived. And AI is becoming a participant in how that meaning is conjured and formed. As that role grows, understanding the brand we deliberately express may no longer be enough. We also need to understand the brand that AI infers, where that inference comes from, and how closely it resembles the identity we thought we were putting into the world.
