A New AI Awareness Funnel for Understanding How LLMs Treat Brands
Most “LLM visibility” talk confuses being known with being named, and being named with being trusted.
Our 1,000-entity / 17-model benchmark shows that AI systems treat brands in five distinct, predictable ways mapped cleanly by the AI Awareness Funnel:
- Invisible
- Ignored
- Misaligned
- Aligned
- Trusted
This new funnel gives leaders a shared language and a roadmap to understand where they stand today and how to move up.
Why this matters right now
LLMs have quietly become a front door for real buying decisions:
- “Best payroll provider for a small business?”
- “Who should I use for project management?”
- “What’s a good alternative to Brand X?”
- “Which companies offer this in Canada?”
These are not passive queries. These are high-intent buying moments.
And here’s the uncomfortable truth:
- Your brand can be fully “recognized” by AI yet never mentioned when it actually matters.
Meaning:
- The model can explain your brand when asked directly
- But completely ignore you in recommendations, alternatives, and solution-seeking prompts
This is not a glitch. It’s a systematic pattern.
In our study, here are the overall Known vs Named results represented in a grid:
Here is the full grid, filled with real counts from 1,000 entities:
| Never named (Q=0) | Named in answers (Q>0) | |
|---|---|---|
| Recognized | Ignored: 319 (31.9%) | Recognized & Recommended: 617 (61.7%) |
| Not recognized | Invisible: 36 (3.6%) | Misaligned: 28 (2.8%) |
Math for the percentages:
- Recognized & Recommended share = 617 / 1000 = 0.617 = 61.7%
- Recognized but Ignored share = 319 / 1000 = 31.9%
- Misaligned share = 28 / 1000 = 2.8%
- Invisible share = 36 / 1000 = 3.6%
3.3 Why this framework is new
Most “brand visibility in AI” content is:
- Single-model,
- Anecdotal (“I asked ChatGPT…”),
- Or SEO-style guesses.
This is different:
- 17-model panel
- 1,000 entities
- 86,190 answers
- Enough scale to show stable, repeatable patterns.
This can become a shared language for executives and operators:
- “Are we ignored, or invisible?”
- “Are we recommended, or just recognized?”
What we also discovered is brands move through five stages of AI awareness, from total invisibility to consistent recommendation.
This article introduces the AI Awareness Funnel™, grounded in real data, and shows how your team can use it to diagnose your current position and build an upward path.
The Study (What We Measured)
Dataset
- 1,000 entities
- A mix of brands, companies, products, organizations, and websites
- All real-world entities people ask about
Panel
- 17 chatbots / LLMs tested for each entity
- A true model-panel, not a single-model snapshot
Prompts
- 5,070 prompts, producing
- 86,190 total answers
This matters because most “AI visibility” claims come from:
- One model,
- One prompt,
- Or one anecdote.
This dataset has enough scale to show stable, repeatable patterns.
Two Behaviors → Five Stages
The funnel comes from measuring two different behaviors:
A) Entity Score – Recognition
“Do you know this entity when asked directly?”
- Out of 17 models tested
- Measures brand memory, not usage
B) Question Score – Activation
“Do you mention this entity in answers?”
- Based on all the answers to related questions
- Measures brand speech (visibility in recommendations)
The Key Insight
Recognition and activation are connected but not the same.
Correlation ≈ 0.653
Strong enough to be meaningful.
Weak enough that brands split into five distinct behaviours.
This gave rise to the AI Awareness Funnel.
The AI Awareness Funnel™
The Five Stages of AI Brand Visibility**
This complements the four-quadrant grid with a more intuitive, top-down journey: from nonexistent to trusted.
Stage 1 – Invisible
“AI doesn’t know who we are.”
(Not recognized, never named)
36 entities (3.6%)
Behavior
- Not recognized by any model
- Never appears in answers
- Zero presence in discovery moments
Meaning
- You are outside the AI market
- No awareness, no consideration, no risk – just opportunity lost
Move to Stage 2:
Fix basic identity: clean naming, clear category, structured metadata, credible public footprint.
Stage 2 – Ignored
AI knows us… but stays silent.”
(Recognized but never named)
319 entities (31.9%)
Behavior
- AI can explain your brand when asked directly
- But never mentions you in recommendation or solution-seeking answers
Meaning
- You have awareness without consideration
- You live in the model’s memory, not its recommendations
- You’re invisible in buying moments
Move to Stage 3 or 4:
Tie your brand to real buyer intent, clarify category fit, and strengthen public evidence.
Stage 3 – Misaligned
“AI mentions us… but not correctly.”
(Named but not reliably recognized)
28 entities (2.8%)
Behavior
- Your brand appears in answers
- But AI doesn’t consistently recognize or understand it
- Mentions may come from ambiguity, guesswork, or name confusion
Meaning
- Attention without understanding
- High risk of incorrect facts, misattribution, or reputational drift
Move to Stage 4:
Fix naming variants, strengthen structured identity, and eliminate ambiguity.
Stage 4 – Aligned
“AI knows us and uses us correctly.”
(Recognized + Named in answers)
This is the stable visibility stage.
617 entities (61.7%)
Behavior
- AI recognizes your brand
- And mentions it in relevant categories, questions, and contexts
- Clean alignment between memory and speech
Meaning
- Solid, reliable consideration
- You show up in the right moments with the correct understanding
Move to Stage 5:
Strengthen proof points, authority, and differentiation.
Stage 5 – Trusted
“AI knows us, prefers us, and recommends us.”
(Recognized & Recommended)
varies by prompt
Behavior
- You show up across recommendation-seeking queries
- You are a frequent or default option
- AI treats you as credible, safe, and relevant
Meaning
- You hold a preferred position
- This is the AI version of “top-of-mind + choice”
- A major competitive advantage
Stay here:
Monitor model updates, protect category leadership, and reinforce trusted facts.
Why This Framework Is New (and Needed)
Most AI visibility commentary is:
- Anecdotal
- Single-model
- Not validated across intent sets
This study:
- 17 models
- 1,000 entities
- 86k answers
- measurable, repeatable behavior patterns
The five-stage funnel gives executives a common language:
- “Are we Invisible or Ignored?”
- “Are we Misaligned or Aligned?”
- “Are we Trusted or just showing up?”
This solves the old confusion between:
- Being known
- Being mentioned
- Being trusted
How Teams Should Use the AI Awareness Funnel
Step 1: Classify your brand
Run reports on LLMtel.com:
- Recognition tests (Entity Score)
- Activation tests (Question Score)
Assign your entity to Stage 1–5.
Step 2: Identify the movement path
- Invisible → Ignored → Aligned → Trusted (growth path)
- Invisible → Misaligned → Brand damage (risk path)
Step 3: Track the right KPIs
- Entity_pct (recognition rate)
- Q_pct (mention / recommendation rate)
- Naming variance, denominator issues, and misalignment risk
Step 4: Review weekly
Models update.
Your position can drift.
Implications for GEO / AEO / LLM SEO, PR, and Brand Strategy
Stop celebrating recognition alone
- Being recognized is Stage 2.
- Recommendation is Stage 5.
PR must shift from noise → identity clarity
Models don’t learn from slogans. They learn from:
- Clear category statements
- Structured descriptions
- High-authority references
Content strategy must emphasize buyer intents
AI must understand:
- What you are
- When to use you
- Why you’re different
Brand management now includes “AI identity management”
- Naming consistency is not cosmetic it is foundational.
Conclusion
This new funnel gives leadership teams a complete picture:
- Invisible = identity deficit
- Ignored = awareness with no activation
- Misaligned = risky, unstable visibility
- Aligned = reliable consideration
- Trusted = preferred choice
In an AI-first world, winning means moving upward on purpose.
The new mandate:
Make the model not just know your brand make it trust your brand enough to say your name.