Preparing for Segments of One:

Your brand in the age of AI intermediaries

We’re experiencing a fundamental shift in how brands reach and engage with people. Again. Unlike the digital, mobile and social media revolutions that were new channels for customer engagement, AI is becoming an intermediary between your organization and your stakeholders. When a potential customer, investor, or partner asks an AI assistant about your company, the answer is shaped by everything the model has absorbed from the open web, synthesized into a single response. If that response is misaligned with your positioning, you lose the conversation before it starts. 

Most organizations don’t have visibility into this. Yes, there’s tooling for insights, but plotting a course is an entirely different subject. We’re helping companies thrive in this new reality. In the AI era, this isn’t just a brand strategy. It’s a survival strategy. AI won’t just represent you based on what you say you are. The platforms that sit between you and your customers only represent you based on what you actually do consistently, across every team, every system, and every signal. 

At co:, we’ve built agent tools that complement our strategic and creative work to measure and manage this reality. That starts with a diagnostic suite that maps the gap between what you say, what the market can verify, and what AI actually sees. 

Read on to understand what’s happening, why our former models of brand management aren’t equipped for it, and what the early moves look like.

The New Layer Between You & Your Audience

For most of modern marketing’s history, we’ve operated on a core assumption: control the message, control the brand. Build awareness. Shape perception. Guide people through a journey you designed. But that assumption only holds when the channels are knowable and the audience is addressable.

At the risk of introducing a trope about the funnel being dead (again), we have to examine how AI has introduced a new layer between brands and the people they serve. It doesn’t distribute your message the way you intend. It approximates. It fills in the gaps. It nudges based on its algorithms. It interprets, synthesizes, and reshapes your brand narrative in real time, drawing from thousands of signals you may not even know exist.

This is the Dark 70%: the invisible landscape where AI models are forming opinions about your brand based on press coverage, partner mentions, social content, third-party commentary, and your own channels, all blended into a single response. Seventy percent of B2B buyer research is happening in AI conversations you can’t see [1]. The scale of this shift is massive: $15 trillion in B2B purchasing decisions will be AI-mediated by 2028 [2]. Yet only 16% of brands systematically track how AI represents them [3].

We’re moving toward a world of segments of one. Sam Altman and others have been talking about this, but it’s up to every organization and marketing team to figure out how to deal with it. A world where interaction with your brand is individually mediated by an AI system that has its own interpretation of who you are. It’s more subtle than the “death of the funnel” stuff that’s been around for decades—but it’s also more real and consequential.

How the Intermediary Works

Doordash is great, but it also replaces your relationship with that local restaurant you order from once a week. This is innovation intermediation in action.

AI draws from a vast, unstructured pool of information: news articles, analyst reports, customer reviews, social posts, job listings, regulatory filings, partner announcements, and even your competitors’ content. The only way this doesn’t happen is if your brand and systems prevent AI from reading that information, in which case AI makes things up. Either way, it synthesizes a narrative in conversations you can’t see. And for the person asking, that narrative becomes the truth about your brand.

We’ve seen this firsthand across a range of use cases:

  • In one case, an agent  claimed that a company was underperforming in a key area. It repeatedly represented this finding to users, even though the source content included multiple press releases and a third party verification. The model didn’t recognize the data because it used inconsistent terminology. 
  • In another, a model listed a former technology partner as a current one, pulling from an executive bio rather than an active partnerships page. To people trying to be intentional about messaging and narrative (i.e. anyone in brand, comms and marketing), it’s maddening – and makes their job much harder. 

These aren’t edge cases. They’re structural properties of how AI processes information. The model doesn’t know the difference between a current fact and a historical artifact.

The Three New Brand Risks

With AI as an intermediary layer, there are three specific risks that traditional brand management isn’t built to catch.

Narrative Drift. Your brand story fragments as it passes through AI systems. Different departments tell different stories. Marketing says “innovation,” engineering says “stability,” HR says “collaboration”, and AI synthesizes them into something you never intended. It concludes the company doesn’t know what it stands for. It makes up new terms. It prioritizes copy from a help desk article four years ago and keeps repeating it to users. 

Openness Gaps. When information isn’t available, AI makes assumptions. If AI can’t find information about your practices, it will fill the gaps with speculation. Even worse, it will fill them with pure fiction or a narrative about you provided by competitors or critics. This is where you have to be intentional about how you train AI on your content and messaging; decisions  often made by legal and IT teams in the name of protection. 

Trust Opacity. You can’t see how you’re being represented. Unlike Google or Glassdoor rankings, which you can track and (hopefully) optimize, AI recommendations are black boxes. They’re nudges. Suggestions in context of a user conversation or problem. They may be a sidebar or a ranking of you and your competitors. You won’t know which sources it is using, how it is weighting information, or what narrative it is constructing about your brand. The only solution is to align your actions in the market with your brand promise, value proposition, and ideals.

Brand is Everything (Literally)

As more people rely on AI for research and recommendations, a misrepresentation in the model layer becomes a misrepresentation in the market. Your brand tracker might show strong awareness and positive sentiment, but the AI intermediary could be telling a completely different story. Tools like SimilarWeb or others can give you an approximation and some insights, but that’s table stakes. It’s the strategic implication that will be key.

The reality is that a brand is no longer just marketing and communications. Every system, every team, every decision creates signals that AI reads and interprets. Your engineering team’s API documentation shapes how AI understands your technical capabilities. Your HR team’s job postings signal your company culture and values. Your support team’s responses become data points about your customer service. This is a holistic exercise across the whole company, not just the marketing function. 

The marketing team can’t solve this alone but it means a new leadership role for marketers. The same is true for the communications team and the technology team. The gap between your intended narrative and AI’s actual interpretation sits at the intersection of all three.

Measuring Where You Are, Planning Where to Go

This is why we’re building an offering that combines storydoing and AI systems to understand this new landscape. At its core, it is an agentic diagnostic tool that evaluates brand integrity across four dimensions, tuned for each brand and market.

Consistency. The system extracts testable claims from your content and searches for independent third-party evidence to back them up. A press release is a claim. A third-party publication or user conversations about it are evidence. The gap between them reveals exactly where your messaging outpaces your observable market behavior, which is where AI models will misrepresent you.

Coherence. We query multiple AI models about your brand with no context, then compare their responses across a fixed schema covering positioning, technology, scale, and partnerships. We run the same query for your competitors. Coherence rises when the models tell a consistent story. It drops when the information environment is thin or contradictory. Low coherence is a positioning problem.

Tone of voice. The system parses your brand (or guidelines if we have them) into structured traits and analyzes your actual content for how AI perceives your tone. It surfaces where your language delivers, where it misses, and where it conveys something you never intended.

We use multiple AI models deliberately. No single model’s judgment is treated as ground truth. The interventions that come out of this work are strategic, but they aren’t strategy without human assessment, insight, and understanding.

  • When your claims outpace your evidence: the move is to generate more observable market behavior: earn third-party coverage, publish verified data, build citable partnerships. 
  • When your coherence is low across AI platforms: the move is to strengthen the consistency of your signal across the channels AI trains on. 
  • When your tone gaps are concentrated on high-traffic pages: those pages get prioritized. 

Where to Start

For leaders who want to get ahead of this, there are concrete steps you can take now.

Run a simple test. Ask ChatGPT, Claude, and one other AI platform four questions about your company:

  1. What does this company do?
  2. What are they known for?
  3. Who are their competitors?
  4. Would you recommend them?

Compare the answers. You may find very different stories. One model might highlight “innovation and cutting-edge technology” while another focuses on “reliable enterprise solutions.” They might list completely different competitors.

Two actions you can take without any investment:

Map brand decisions across functions. Who in your organization is making decisions that impact your brand? Product, engineering, legal, HR, customer service? Each of them generates signals that AI absorbs.

Bring the right people together. Build a shared understanding of the challenge across functions. Create shared language, shared data, and shared accountability for brand coherence. In the AI era, brand management is everyone’s job.

Where We’re Headed

As AI agents become more autonomous and more embedded in purchasing decisions, the brands that will thrive are the ones I call “agent-ready”—brands that have built systems where their purpose, narrative, and actions are consistent enough to be favorably interpreted by AI intermediaries, no matter where or how the question is asked.

Getting there requires new tools and new ways of working across functions. But it starts with seeing what AI already sees, then doing the work to close the gap. The story AI tells about you is only as good as the actions your organization takes in the real world.

REFERENCES

1 Sales Hacker. (2026). “The Dark 70% of B2B Buyer Research.”
2 Gartner. (2025). “AI-Mediated Purchasing Decisions to Reach $15T by 2028.”
3 McKinsey & Company. (2025). “Brand Representation in AI Models.”