The Traditional Marketing Funnel is Losing Its Foundation
For over thirty years, the digital marketing industry has operated on a top-down architecture. The strategy was linear and seemingly logical: start with broad awareness, cast the widest possible net, and gradually nurture prospects down through a narrowing funnel. This model assumed that if you could simply get in front of enough people, the sheer volume of the top-of-funnel (TOFU) would eventually yield results at the bottom.
In the era of broadcast media, this made perfect sense. In the early era of search engines, it remained largely effective. However, as we transition into an environment dominated by Artificial Intelligence, large language models (LLMs), and autonomous agents, the top-down approach isn’t just inefficient—it is fundamentally flawed. AI-driven systems do not evaluate brands from the top down; they build their recommendations from the bottom up.
Search engines, assistive engines like Perplexity, and AI agents like Claude or ChatGPT must first understand who you are before they can determine if you are credible. They must verify your credibility before they even consider recommending you to a user. If you continue to pour your budget into top-down awareness without establishing this foundational understanding, you are effectively building a house on sand. The AI agents will have no structural foundation to attach your brand to, leaving your marketing efforts invisible to the very systems that now mediate the customer journey.
The 128-Year-Old Model Meets a Structural Break
The concept of the acquisition funnel isn’t new. It was formalized in 1898 by Elias St. Elmo Lewis. For 128 years, every marketing department on the planet has leaned on some variation of his AIDA model: Awareness, Interest, Desire, and Action. While the channels have shifted from newspapers to radio to social media, the direction remained the same: reach first, relationship second, commitment third.
In 2002, Philippe Lanceleur famously described the early web as a shop in the middle of a field. You couldn’t just build it and hope for visitors; you had to go where people were and lead them back to your shop. Awareness remained the prerequisite. However, the introduction of the Knowledge Graph by Google in 2012 signaled the beginning of a shift. Suddenly, the machine began forming its own opinions about brands independently of user queries. The machine started drawing its own maps and building the roads for the users.
With the rise of agential AI, we are seeing the first genuine structural break in marketing strategy since the 19th century. While the user experience still looks like a traditional funnel—they hear about you, evaluate you, and then decide—the strategy to get the machine to surface your brand must be flipped. To the AI, your brand’s understanding and reputation are the prerequisites for awareness, not the other way around.
Understanding the Bi-Directional Funnel
In the modern tech landscape, the acquisition funnel now runs in two opposite directions simultaneously. To navigate this, marketers must understand that while the human user moves from the top down, the machine moves from the bottom up.
The machine’s internal logic follows this sequence:
1. Understandability (The Foundation)
Does the machine know who you are? This is the bottom-of-funnel (BOFU) layer for the AI. If the engine cannot resolve your brand as a specific entity with defined attributes, it cannot process you. This is the moment of identity. Without a clear entity node, the machine has nothing to recommend.
2. Credibility (The Evaluation)
Does the machine trust what you do? Once the AI identifies you, it works upward to assess your reputation. It looks for signals of Experience, Expertise, Authoritativeness, and Trustworthiness (E-E-A-T). If the machine understands who you are but finds your competitor more credible, the agent will act in favor of the competitor every time.
3. Deliverability (The Recommendation)
Will the machine proactively recommend you? This is the top-of-funnel (TOFU) layer for the AI. Only after understanding and trust are established will the engine “deliver” your brand to a user who hasn’t specifically asked for you. This is the ultimate zero-sum moment: the recommendation that happens in a private conversation between a user and an AI, for a prospect you didn’t even know was in the market.
How Top-Down and Bottom-Up Coexist
It is important to note that the traditional top-down funnel hasn’t disappeared; it has simply been augmented. You can still build top-down awareness in channels you control entirely—such as paid media, direct mail, or broadcast advertising. You can buy attention and pull people toward a decision.
However, within organic ecosystems—where AI engines and agents act as mediators—you must build from the bottom up. Every algorithm and assistive agent now operates on brand signals rather than just volume. Reach on social media is increasingly dictated by brand recognition and topical authority. In this environment, the machine-built roads to your “shop in the field” are constructed from brand understanding outward to awareness.
If the machine doesn’t understand you, it won’t recommend you. If it doesn’t trust you, it will hedge its answers or stay silent. This is a mechanical reality of how modern AI infrastructure is built.
The AI Engine Pipeline: The 10 Gates to Success
Winning the AI recommendation requires passing through a series of “gates” within the AI engine pipeline. This journey from being discovered to being “won” is a 10-stage process that highlights why the bottom-up approach is mandatory.
The first five gates are infrastructure-based. They involve the machine’s ability to access, store, and classify your content. This is the “Annotation” phase. If you fail here, you don’t even exist in the machine’s world. However, from Gate 6 (Recruitment) onward, the engine begins comparing you to every other alternative in the market.
The pipeline culminates in the “Display” gate, where the machine makes a final judgment. It is here that the Understandability, Credibility, and Deliverability (UCD) layer becomes visible to the user. If the AI is not fully convinced of your brand’s identity and merit by Gate 8, you will lose at Gate 9 (the “Won” stage).
The Business Case for UCD: The Three Taxes of AI Failure
To explain this to a non-technical audience or a C-suite executive, it helps to think of the major AI platforms—Google, ChatGPT, Claude, Siri, Alexa—as a global salesforce. These “employees” work 24/7. They are either selling your products or they are selling your competitors’ products. If you haven’t trained them properly, you are paying what we call the “Three Taxes.”
The Doubt Tax (Understandability Failure)
This occurs when a prospect is ready to buy and asks the AI about you, but the machine cannot confirm who you are or what you offer. You will see responses like “This brand appears to offer…” or “I don’t have enough information about…” This tax is a direct result of weak entity resolution and poor foundational data.
The Ghost Tax (Credibility Failure)
This happens when an AI is asked to compare options in your industry. If the machine’s confidence in your credibility is lower than its confidence in your competitor, you are simply left off the shortlist. You become a “ghost” in the evaluation process. You were never even considered, and you’ll never know you lost the lead.
The Invisibility Tax (Deliverability Failure)
This is the cost of never being mentioned to prospects who are researching a problem you solve. When a user asks an AI for a solution to a pain point, the machine only advocates for brands it already understands (U) and trusts (C). If you haven’t built those layers, you remain invisible during the research phase.
The Funnel as a Guided Sequence in AI
In the traditional search era, Google provided a list of links (the SERP), and the user did the hard work of navigating the funnel themselves. They clicked, compared, and moved from awareness to decision through their own browsing habits.
In the AI era, the LLM takes over this navigation. Through a process of “fan-out” or “cascading” queries, the AI reasons about the user’s intent and anticipates the next logical step. The AI isn’t just answering a question; it is shaping the entire acquisition journey. It defines the sequence of follow-up questions and sets the stage for the user’s next move.
As a brand, your job is to train the machine’s expectations. You must publish the “logical bridges”—content that explains why, if a user is considering X, the natural next step is Y. By providing this evidence and corroboration across multiple trusted sources, you encourage the machine to treat these bridges as settled facts. When the machine predicts the user’s next step, your brand becomes the natural, logical destination it reaches for.
How to Start: The Brand SERP and AI Résumé
If you are wondering where to begin your bottom-up strategy, look at your Brand SERP (what Google shows when someone searches your name) and your AI Résumé (how a conversational agent describes you). These are your diagnostic instruments.
If the AI gets your basic facts wrong, you have an Understandability problem. You need to focus on your “Entity Home”—a single, authoritative source of truth for your brand, supported by clean structured data and schema markup.
If the AI’s description of you is unconvincing or ignores your key strengths, you have a Credibility problem. You need to focus on off-site signals: earned media, third-party reviews, and co-citations from authoritative sources the machine already trusts.
If the AI understands and trusts you but never suggests you for broader industry queries, you have a Deliverability problem. You need to create content that serves as “proof” of your expertise, ensuring the machine treats your brand as the reference option for your category.
The 15-Stage Play: Beyond the Acquisition Funnel
While we focus heavily on the acquisition funnel because that is where the “Won” moment occurs, it is actually just one act in a much larger 15-stage play. The AI’s confidence in your brand is not just built before the sale; it is reinforced after the sale.
The gates following the “Won” stage—Onboarded, Performed, Integrated, Devoted, and Codified—are where the real business value lies. Every satisfied client creates signals that feed back into the machine’s knowledge base. When a client’s success is documented and recognized, it strengthens the machine’s confidence for the next prospect. This creates a flywheel effect: successful outcomes lead to higher machine confidence, which leads to more recommendations, which leads to more revenue.
In an AI-mediated world, marketing is no longer just about shouting the loudest at the top of the funnel. It is about building a foundation of understandability, earning a reputation for credibility, and providing the machine with the confidence it needs to advocate for you. The funnel has flipped, and those who adapt to this bottom-up reality will be the ones the machines choose to recommend.