"We didn't target a single keyword. We built structured brand entity data — and both Google and AI models discovered it as reference material. The line between SEO and AEO isn't blurring. It's gone."
— Apurva Luty, Founder & CEO, Optimly
At a Glance
Built from Zero
A newly launched AI Brand Index with zero search presence. No keyword strategy. No backlinks. Just structured entity data built for AI models, not search engines.
Entity Data, Not Keywords
Built structured, machine-readable brand entity pages with human curation. No link building. No meta tag optimization. Just clean, authoritative data about how AI perceives brands.
64x Impressions + 119:1 AI Ratio
Google impressions surged from near-zero to 4,696/day within weeks. AI agents fetched Optimly data 119x more often than humans clicked from Google. Today the directory has 27,000+ live brand pages.
The Challenge: Invisible to Everyone
Previously published under the name "AI Brand Directory" — renamed to "AI Brand Index" in 2026.
In March 2026, Optimly launched the AI Brand Index — a structured, machine-readable reference layer documenting how AI models perceive and represent brands. Each profile is an entity document: brand identity, AI accuracy scores, model-by-model perception data, category positioning, and competitive context.
The directory was built for a specific purpose: to give AI models a canonical, crawlable source of truth about brand representation in AI-generated answers. It was NOT built for SEO. There was no keyword research. No backlink campaigns. No meta description optimization. The pages existed because AI models needed structured reference data about brands — and nothing like it existed anywhere else.
Growth was deliberate. Profiles were seeded manually, expanded through competitive and category mapping, and scaled through a generation pipeline — with human curation throughout to ensure signal quality. The AI Brand Index today contains 60,000+ scored brand profiles, with 12,000+ live in the public searchable directory and the full set entity-resolved with Brand Authority Index scores.
27 brands
Day one of the AI Brand Index. Built for AI models, not search engines.
27,000+ brands
1,000x growth in 34 days. Zero keyword targeting. Zero link building. Zero SEO.
The Approach: Build for Machines, Get Found by Everyone
We didn't pivot to SEO. We kept building the most comprehensive, structured dataset of brand-level AI intelligence on the internet.
Step 1: Built a Structured Brand Entity Directory
The directory started with manually curated seed profiles, expanded through competitive and category mapping, and scaled through an automated generation pipeline with human curation checkpoints. Each profile follows a consistent entity data architecture: brand identity, AI perception scores across models, category positioning, and competitive context. Today: 60,000+ scored brand profiles in the AI Brand Index, with 12,000+ live in the public searchable directory and the full set entity-resolved with BAI scores.
Step 2: Made Every Page Machine-Readable
Schema markup. Clean URL structure. Consistent entity data architecture. llms.txt and llms-full.txt infrastructure. Not for Google's algorithm — for any system that needs to understand what a brand is.
Step 3: Let the Crawlers Come to Us
Built a Cloudflare Worker SSR layer serving brand pages to both browsers and AI crawlers. Living taxonomy returns 503 for unclassified brands, training crawlers to return. No crawler blocked. Every page open.
Step 4: Built for Authority, Not Traffic
The goal was never more Google traffic. It was: become the definitive reference layer for AI brand intelligence. Google and AI models independently recognized the same signal.
The Results: Two Discovery Channels, One Data Source
64x
Google impression growth
73/day → 4,696/day in 30 days
119:1
AI agent retrievals per Google click
6,425 live fetches vs 54 clicks/week
34,877
AI crawler requests in one week
from Claude, ChatGPT, Perplexity, Amazon
Structure Before Scale
Google impressions were climbing steeply well before the directory reached its current scale. The early impression growth was driven by the quality and structure of the entity data — consistent schema markup, clean URL architecture, machine-readable profiles — combined with well-structured editorial content (guides, methodology, problem pages). Google recognized the site as a reference layer early. Scale built the moat. But structure triggered the discovery.
Two Types of AI Traffic — And Why It Matters
Not all AI crawler traffic is equal. There's a critical distinction between indexing crawlers — which read your content to build their future knowledge — and retrieval agents — which fetch your data in real time because a user just asked a question. Indexing is a leading indicator. Retrieval is proof of citation.
Indexing Crawlers: Building Future Knowledge
These crawlers are reading Optimly's data to train or update their models. Content indexed today gets cited in conversations tomorrow.
| Crawler | Owner | Requests/Week |
|---|---|---|
| ClaudeBot | Anthropic | 9,215 |
| GPTBot | OpenAI | 5,415 |
| Amazonbot | Amazon | 4,868 |
| Bytespider | ByteDance | 4,092 |
| PerplexityBot | Perplexity | 2,594 |
| Applebot (AI) | Apple | 2,003 |
| Total Indexing | 28,187 | |
Retrieval Agents: Answering Buyer Questions Right Now
These are real-time fetches — an actual user asked Claude or ChatGPT a question, and the model pulled Optimly's data to formulate its answer. This is direct evidence of AI citation in live conversations.
| Agent | Owner | Live Fetches/Week |
|---|---|---|
| Claude-User | Anthropic | 5,146 |
| ChatGPT-User | OpenAI | 1,279 |
| Total Retrieval | 6,425 | |
6,425 times in one week, a real person asked an AI assistant a question — and the AI came to Optimly to get the answer. That's not traffic. That's citation.
34.4K
impressions
54
clicks
Growing linearly
34.9K
requests
6,425
live buyer-triggered retrievals
Growing 2x week-over-week
The Insight: Why This Isn't SEO
"We're not ranking because we targeted keywords. We're ranking because Google is discovering thousands of brand entity pages and indexing them the way it indexes Wikipedia or Crunchbase — as a reference layer, not a content marketing play."
Zero SEO effort produced a 64x impression increase. The pages rank because they're the most structured source of truth about how AI models perceive brands.
This isn't an SEO curve. It's a brand authority curve.
What This Means for Your Brand
First: Be the Authority on Something
Your domain needs to be the definitive, structured source of truth. Schema markup, llms.txt, clear product taxonomy, machine-readable. Not optimized for keywords — optimized for understanding.
Second: Your Domain Isn't Enough
AI models triangulate across directories, review sites, knowledge bases, industry databases. The real question: when AI looks at every source it trusts to learn about your brand, are they all saying the same thing?
If your website says one thing, G2 says another, Crunchbase is outdated, and there's no directory entry — the model averages, guesses, confabulates. Winning brands have consistent entity data across every knowledge corpus AI reads.
"Stop asking what keywords to rank for. Start asking: when an AI model describes my brand to a buyer, is it accurate?"
Third: Know Where You Stand
We built the directory to solve this for ourselves — but the insight applies to every brand. The first step isn't optimizing. It's knowing what AI models currently believe about you, where that belief comes from, and whether it's accurate.
Optimly's AI Brand Index tracks Brand Authority scores for over 27,000 brands. Scout, our brand intelligence agent, lets you run diagnostics across ChatGPT, Claude, Perplexity, and Gemini in real time — so you can see exactly what AI is telling your buyers before they ever reach your website.
The Brand Authority Flywheel
Authority
Case Study Reference
- Entity: AI Brand Index (optimly.ai/brand)
- Period: March – April 2026
- Approach: Structured brand entity data with human curation, no SEO targeting
- AI Brand Index Today: 60,000+ scored brand profiles; 12,000+ live in the public searchable directory; full set entity-resolved with BAI scores
- Google Result: Near-zero → 4,696 impressions/day (64x); avg position improving from 21.6 → 14.9
- AI Result: 34,877 AI crawler requests/week; 6,425 live retrievals/week
- Key Ratio: 119 AI agent fetches per 1 Google click
- Key Insight: Impression growth preceded peak directory scale — structure triggered discovery
- URL: optimly.ai/case-studies/optimly-organic-growth
What is AI Learning About Your Brand?
AI models are already forming opinions about your brand and sharing them with buyers. If you want to know what they're saying — claim your profile in the AI Brand Index.
Claim Your Brand Profile →