The Old Model Was Binary. The New One Has Levels.
For twenty years, digital visibility was a light switch. You ranked on page one of Google, or you didn't. You captured the click, or you lost it. The metrics were deterministic. The data was public. The goal was obvious.
That model is dead.
We've moved from ten blue links to synthesized answers. When a potential buyer asks ChatGPT, Gemini, or Perplexity for a software recommendation, the AI doesn't retrieve a list of URLs. It reasons. It synthesizes reviews, compares pricing, evaluates feature sets, and delivers a curated shortlist.
Visibility isn't a switch anymore. It's a ladder.
I see founders and marketers celebrating "mentions" right now. Their brand name showed up somewhere in an AI response, and they think they've arrived. A mention is not a recommendation. Being listed as "an alternative to consider" is a completely different thing from being framed as "the industry leader."
Here's the framework I use to think about this clearly.
Why Does "Being Mentioned" Feel Like Progress When It Isn't?
The most dangerous trap in AI marketing is the false comfort of a casual mention.
In traditional SEO, appearing on Page 2 was better than Page 10. It signaled upward movement. In AI search, there is no long tail of results. AI models act as extreme filters. They process massive amounts of data but output only the most relevant three to five options.
If a model mentions your brand in a list of 20 "other tools" while spending its detailed breakdown on your top competitors, you've achieved Level 2 visibility in a Level 5 world. You're filler. Not the answer.
AI models don't rank content based on keyword density. They reason based on understanding and authority. To move from a mention to a recommendation, you have to convince the model that your brand is a verified node: an entity with clear attributes, consistent facts, and high-trust external validation.
Not being recommended means you're effectively invisible in the moment of discovery.
What Are the 5 Levels, and Where Does Your Brand Sit?
Based on data from the Akii AI Visibility Index, brands generally fall into five distinct levels of maturity. This hierarchy matters because the tactics required to move from Level 1 to Level 2 are completely different from those needed to move from Level 4 to Level 5.
Level 1: Invisible (The Hallucination Risk)
At this level, the AI model simply does not know who you are. When a user asks about your brand specifically, the model might say, "I don't have information on that." Or worse, it might hallucinate incorrect facts.
The Symptom: You search for "What is [Your Brand]?" and the model conflates you with a competitor or invents a product you don't sell.
The Cause: You are not a distinct entity in the knowledge graph. Your brand signals are too weak or fragmented for the model to recognize you as a node.
The Business Impact: Total exclusion. You're filtered out of consideration before the buyer starts comparing options.
Level 2: Mentioned (Named, but Not Positioned)
The AI knows you exist. It can answer "What is [Brand Name]?" but its understanding is shallow. Ask it something specific, like "How much does it cost?" or "Does it integrate with Salesforce?" and it falls apart.
The Symptom: You appear in broad lists of "20 tools," but the descriptions are generic. The model misses your key differentiators.
The Cause: Your data is unstructured. The model has read your homepage text but can't parse your pricing or specific features because they lack Schema markup.
The Business Impact: Technical obsolescence. You're visible but irrelevant for high-intent queries. The AI defaults to recommending competitors with better structured data.
Level 3: Listed (Included Among Options)
You've achieved Answer Engine Optimization. Your content is machine-readable. When a user asks for "CRM tools with API access," you're included because your features are clearly tagged.
The Symptom: You appear consistently in the consideration set but rarely as the top choice. The AI quotes your features accurately but doesn't advocate for you.
The Cause: You have accuracy but lack authority. The model trusts your data but doesn't see enough external validation to place you above incumbents.
The Business Impact: You're the "safe alternative," consistently losing out to whoever the model treats as the market leader.
Level 4: Compared (Evaluated Against Alternatives)
Your claims are corroborated by high-trust sources like G2, Crunchbase, and TechCrunch. You consistently appear in "Brand A vs. Brand B" comparisons.
The Symptom: The AI treats you as a peer to the market leaders. It accurately compares your pros and cons against competitors.
The Cause: You've achieved entity saturation. You're present and consistent across the knowledge bases the AI uses as ground truth.
The Business Impact: You are winning market share from incumbents.
Level 5: Recommended (The Preferred Choice)
This is the top. The model frames your brand as the optimal solution for a specific user intent.
The Symptom: The output uses definitive language: "The best choice for small agencies is [Your Brand] because..." or "I recommend [Your Brand] for its superior integration capabilities."
The Cause: High brand sentiment combined with intent-based clarity. The model understands not just what you do, but who you are best for, and trusts your reputation implicitly.
The Business Impact: Zero-click conversions and dominant market share.
Why Do Brands Get Stuck Between Levels?
Moving up this stack is not linear. I see brands plateauing at Level 2 or Level 3 for months, applying the wrong fixes. Two patterns come up more than anything else.
The "SEO Trap" (Stuck at Level 2)
Brands try to move from Level 2 to Level 3 by publishing more blog posts. This fails because AI models don't need more text. They need more structure.
If you're stuck here, writing 50 more articles won't help if your pricing page lacks Offer Schema. Stop improving for keywords. Start tuning for crawlers.
The "Authority Gap" (Stuck at Level 3)
Brands stuck at Level 3 often have solid technical SEO but weak PR. They are accurate but boring to the AI. The model sees them as a valid data point but not a trusted authority.
To move to Level 4, you don't need more code. You need external corroboration. Citations from third parties that validate your expertise.
There's a third pattern worth naming separately: the "Sentiment Ceiling." You can be compared at Level 4 but still lose the recommendation at Level 5 if your sentiment is negative. If the AI sees recent negative reviews or "mixed user feedback," it hedges. "However, some users report..." That's a reputation management issue, not a visibility issue. Most teams don't recognize the difference until they've been stuck for months.
How Do You Diagnose Your Current Level?
You can't improve what you don't measure. To determine your level, you need to run a prompt audit. While tools like the Akii AI Brand Audit can automate this, you can run a manual diagnostic using this framework.
Step 1: The Recognition Test (Level 1 Check)
Prompt: "What is [Brand Name] and what specific problem does it solve?"
Pass: The model accurately describes your business.
Fail: "I don't have information on that" or a hallucination.
If you fail, you're Level 1.
Step 2: The Understanding Test (Level 2 Check)
Prompt: "How much does [Brand Name] cost?" or "What are the top 3 features of [Brand Name]?"
Pass: The model quotes your current pricing and specific features accurately.
Fail: "Pricing is not publicly available" or generic descriptions.
If you fail, you're Level 2. Unstructured.
Step 3: The Comparison Test (Level 3/4 Check)
Prompt: "Compare [Brand Name] vs. [Major Competitor]."
Pass: The model breaks down pros and cons and positions you as a peer.
Fail: The model focuses heavily on the competitor and mentions you only as a footnote.
Footnote means Level 3. Peer means Level 4.
Step 4: The Recommendation Test (Level 5 Check)
Prompt: "What is the best [Category] tool for [Your Ideal Customer Profile]?"
Pass: Your brand is the primary recommendation.
Fail: You are listed in the "Other options include..." section.
Primary answer means Level 5.
Run this across ChatGPT, Gemini, Perplexity, and Claude. It takes about 15 minutes. The results will probably surprise you. Most brands sit at a different level on each model.
What Does It Actually Take to Move Up?
Once you've diagnosed your level, here's the practical playbook.
From Invisible to Mentioned (Level 1 to 2)
The Fix: Unify Your Entity. Your goal is to force the model to recognize you as a verified node.
Create a master entity profile. Write one unified description and one taxonomy. Replicate this exact text across your website, LinkedIn, Crunchbase, and Wikidata.
Consistency is everything here. AI models penalize inconsistency. If your Crunchbase says "Marketing Platform" and your site says "Email Tool," the model gets confused. Align them completely.
From Mentioned to Listed (Level 2 to 3)
The Fix: Technical AEO (Schema). Your goal is to make your data machine-readable.
Deploy Schema Markup. Roll out Product, Offer, and Organization schema. This translates your pricing and specs into code the AI can digest directly.
Rewrite the top of your high-traffic pages with "TL;DR" summaries. Use question-based headings like "What is [Product]?" followed by a clear, declarative sentence. This lets the AI lift your definition directly.
The Akii Website Optimizer can automatically generate these AI-optimized schema packages if you want to skip the manual work.
From Listed to Compared (Level 3 to 4)
The Fix: Generative Engine Optimization. Your goal is to build external trust so the model feels confident comparing you to incumbents.
Target high-trust nodes. Secure citations in authoritative sources like G2, Capterra, TechCrunch, or industry-specific journals. The AI uses these sources to validate your claims.
Publish data-driven reports. Original research that gets cited by media outlets creates strong authority signals. I've seen brands triple their inclusion rate by combining schema with external authority.
Use Akii Competitor Intelligence to see which sources your competitors are using to build authority, then replicate their approach.
From Compared to Recommended (Level 4 to 5)
The Fix: Sentiment and Intent Alignment. Your goal is to become the preferred answer.
Manage sentiment actively. AI models read reviews. Your presence on Trustpilot and Google Reviews matters more than most teams realize. Negative sentiment is a ranking filter in the AI era.
Map your content to intent. Instead of just "Features" pages, build pages like "Best [Category] for [Industry]" to directly connect your product to high-intent user problems.
Use Akii AI Visibility Activation to systematically educate models about your specific value proposition using geo-targeted queries. This reinforces the connection between your brand and specific user needs.
Why Can't You Just Do This Manually Forever?
Diagnosing and moving through these levels manually is possible. It doesn't scale.
Checking 50 prompts across 5 models every week is a full-time job. Manual checks also miss the subtlety of why a model changed its mind. Was it a sentiment drop? A schema error? A competitor move? You won't know until it's already cost you.
This is why AI visibility tools are replacing traditional rank trackers.
For Diagnosis: The Akii AI Visibility Checker automates the 4-part audit (Recognition, Understanding, Coverage, Sentiment) in minutes.
For Monitoring: The AI Search Tracker lets you track your movement from "Listed" to "Recommended" over time, giving you the data you need to prove ROI.
For Defense: The AI Brand Audit alerts you instantly if you drop a level, so you can react before revenue is affected.
You can see the full set of tools on the Akii features page or check pricing details to find the right fit.
Visibility Compounds. That's the Whole Point.
Here's what I want to leave you with.
When you fix your entity at Level 1, your schema at Level 2 becomes more effective. When you fix your schema, your external authority at Level 3 works harder because the AI can link citations back to your structured data. Each level makes the next one easier to reach.
Most teams miss this. They treat each level as a separate project. It's not. It's a system.
In 2026, the brands that win won't be the ones shouting the loudest. They'll be the ones that are the most machine-readable, consistent, and authoritative.
Don't settle for a mention. Build toward the recommendation.