What are the 4 levels of a firm's AI Visibility?
The ReBo framework sorts financial firms into 4 levels of AI Visibility: Level 1 — Invisible to AI (the models don't know the firm exists), Level 2 — Partially Discoverable (they find fragments), Level 3 — AI Friendly (they read it, understand it, and start citing it), and Level 4 — AI Reference (the firm gets cited ahead of its competitors). Most Brazilian financial institutions sit at Level 1.
Level 1 — Invisible to AI: what does it mean?
Every piece of the firm's content sits in a format AI crawlers can't read: monthly letters as PDFs, JavaScript-rendered websites, research locked behind a login or a newsletter. GPTBot, ClaudeBot, and PerplexityBot don't execute JavaScript, so as far as they're concerned that content does not exist. When an allocator asks ChatGPT about the firm's market, the firm never shows up in the answer.
Level 2 — Partially Discoverable: what does it look like?
Some content is accessible: a corporate website, a handful of LinkedIn posts. The models find fragments. They know the firm exists, but they have nothing deep enough to cite as a source. In the answers, the firm shows up as a passing mention at best, never as the analysis worth quoting.
Level 3 — AI Friendly: what defines a firm at this level?
The content is structured, semantic, and published on a steady cadence: static HTML with a clear hierarchy, structured data (JSON-LD), FAQs, one dedicated page per question, and visible dates. The models read it, understand it, and start citing it, because the structure hands them question, answer, author, and date with no ambiguity — exactly what a generated answer needs to cite a source with confidence.
Level 4 — AI Reference: how does a firm get there?
The firm gets cited again and again in answers about its market, ahead of its competitors. On top of the technical foundation from Level 3, this takes verifiable authority (E-E-A-T): named authors with credentials, proprietary data, a consistent presence across platforms, and ongoing monitoring of AI answers. This is the level the ReBo method targets for every client.
How do the 4 levels compare?
| LEVEL | WHAT THE AI SEES | OUTCOME IN THE ANSWERS |
|---|---|---|
| 1 · Invisible to AI | Nothing — PDFs and JavaScript are unreadable | Never cited |
| 2 · Partially Discoverable | Corporate fragments only | Passing mentions, no research cited |
| 3 · AI Friendly | Structured HTML, JSON-LD, FAQs, dates | Citations start showing up |
| 4 · AI Reference | Verifiable authority + proprietary data | Cited routinely, ahead of competitors |
Source: ReBo AI Visibility Framework, 2026.
How do I find out which level my firm is at?
Test the questions your clients are already asking the models. The ReBo diagnostic tracks which ChatGPT, Perplexity, Gemini, and Claude answers name your firm, and it comes with the first conversation. For the concepts behind the framework, start with what AI Visibility is and why AI models can't read PDFs.
Frequently asked questions about the framework
Which AI Visibility level are most firms at?
Level 1 — Invisible to AI. Most of what Brazilian financial firms publish sits in PDFs and JavaScript-dependent pages — formats AI crawlers like GPTBot, ClaudeBot, and PerplexityBot can't read.
How long does it take to move from Level 1 to Level 3?
With a structured method, the technical foundation and the first AI-friendly pages go live in 4 to 8 weeks. Steady citation — Level 3 fully established — depends on publishing consistently, and it typically shows up over the months that follow.
Do we need to rebuild our website to move up a level?
Not necessarily. What matters is that every strategic piece of content lives as static HTML, with a semantic hierarchy, structured data, and one dedicated page per question. That layer can go up alongside your current website.
How is the level measured?
We run a set of target questions through ChatGPT, Perplexity, Gemini, and Claude, then record where the firm gets cited, how often, and in what position. That diagnostic is ReBo's first deliverable.