Aimee Q Devlin · Methodology
An AI search visibility framework for founder-led businesses
Most AEO advice starts at the technical layer. It assumes you already know who you are. It optimises the infrastructure before the identity is clear. This is why most AEO work doesn't stick—you're building on ground that hasn't been surveyed.
PRISM is a sequential five-layer framework for making a business visible to AI search engines—ChatGPT, Perplexity, Claude, Google AI Mode, and the agentic systems increasingly researching and recommending on behalf of human buyers. Each layer enables the next. You cannot skip ahead. And it starts where others don't: with the person, not the platform.
Upstream—before the build
The five PRISM layers are a build and audit framework. But they rest on two upstream methodology layers that are not part of the public tool—because they require a conversation, not a crawler.
Excavation
Deep personal discovery: working style, personality, Human Design, brand archetypes, goals, challenges, competitors, market gaps. This is the instruction manual. It determines what gets amplified and what gets retired.
Positioning
The gap you own. Who you're for. What you want to be known for. Every downstream decision is built on this. Most AEO work skips these two steps. That's why it doesn't stick—you can't build infrastructure for the wrong identity and expect it to hold.
Platform is how AI finds you in the first place.
Can the crawlers get in?
Before any AI engine can cite you, its crawler has to be able to access and read your site. This sounds basic. It is not. Most small businesses are accidentally blocking the AI crawlers that matter—through overly broad robots.txt rules written during the 2023–2024 bot-blocking wave, or through JavaScript-rendered sites that return near-empty HTML to anything that isn't a browser.
A site that looks complete and polished to a human visitor can be a blank page to ChatGPT. The crawler arrives, retrieves the raw HTML, finds nothing—because the content loads after JavaScript executes, and AI crawlers don't execute JavaScript. There's no error. No warning. Just absence.
Platform covers bot access (robots.txt directives for GPTBot, ClaudeBot, PerplexityBot and others), HTML-first rendering, sitemap.xml, llms.txt, Core Web Vitals, and IndexNow. It is the most mechanical layer—but it has to be right before anything else matters.
If Platform is broken, nothing downstream exists. AI engines cannot cite what they cannot read.
Once Platform is in place, the next question is does ai know who you are?
Recognition tells AI whose house it just walked into.
Does AI know who you are?
Getting in the door is one thing. Being recognised is another. Once a crawler can access your site, it needs to understand the entity behind it—who you are, what you do, and whether that entity is consistent across the web. This is Recognition.
Entity definition covers Person and Organisation schema, sameAs links that connect your site to trusted third-party profiles, consistent naming across every platform you appear on, and an About page that functions as your entity home—the canonical place where AI comes to understand what you are.
Most small businesses have zero schema, inconsistent naming between their website and LinkedIn and Google Business Profile, and old profiles from past projects or former roles that muddy the entity. AI doesn't benefit of the doubt. Confusion for humans is confusion for robots—it just happens faster and more completely. If a human can't figure out who you are in thirty seconds of googling, AI can't either.
Without Recognition, AI may be able to read your site—but it doesn't know whose it is. You become content without an author. Information without attribution.
Once Recognition is in place, the next question is can ai understand what you know?
Intelligence is your knowledge, made machine-readable.
Can AI understand what you know?
You have knowledge. The question is whether it exists in a format that AI can retrieve, parse, and cite. Most of it doesn't. Most founder knowledge lives in social posts, podcast episodes, email newsletters, and conversations—formats that AI cannot reliably access or attribute.
Intelligence is the content architecture layer: answer-first writing structure, FAQ schema, hub-and-spoke pillar pages, keyword research integrated into content planning, and glossary entries that let you define the terms in your category rather than inheriting someone else's definitions. It prioritises owned-domain content over rented social platforms, because what lives on your domain is indexed and attributable in a way that a LinkedIn post simply is not.
The goal is not to dumb things down. It is to translate without diluting—to take what you know and make it legible to a machine that is very good at finding structured answers and very bad at inferring expertise from conversational prose.
Without Intelligence, you may be recognised—but AI has nothing to cite. It knows who you are but can't explain what you know.
Once Intelligence is in place, the next question is does anyone else say this about you?
Signal is the world independently confirming what you claim.
Does anyone else say this about you?
AI validates brands through third-party sources before it trusts what appears on a website. This is not a new behaviour—it mirrors how trust has always been built. What you say about yourself matters less than what others say about you, in places that AI treats as authoritative.
Signal is the off-site layer: podcasts and guest articles that exist as indexed content, structured case studies with named outcomes (not vague testimonials), community presence, directory listings, LinkedIn treated as a trusted entity source, and reviews translated into citable content rather than locked inside star-rating widgets.
For founder-led businesses, Signal isn't a PR campaign. It's strategic placement—the right piece in the right room. A single well-placed interview in a relevant publication, properly structured for AI retrieval, is worth more than a hundred social posts. The question is not how much signal you have. It's whether the signal you have is in places AI looks.
Without Signal, you're asking AI to take your word for it. AI doesn't. Absent from trusted third-party sources means absent from recommendations.
Once Signal is in place, the next question is is your infrastructure getting smarter—or just older?
Momentum is the compound interest on everything above it.
Is your infrastructure getting smarter—or just older?
Infrastructure built once and left alone decays. The crawlers change. The AI engines update what they retrieve and how. Competitors build their own infrastructure. New signal sources emerge. Content that was well-structured two years ago may no longer match how AI is parsing answers today.
Momentum is the retainer layer—phased execution, content calendar, page iteration based on what's actually appearing in AI results, bot analytics that track which crawlers are visiting and what they're retrieving, and commercial query monitoring across ChatGPT, Perplexity, and Claude to see where you appear and where you don't.
This layer runs on both data and instinct. Some of what compounds is measurable—rankings, citation frequency, traffic from AI referrals. Some of it is pattern recognition built from watching many sites change over time. The question this layer answers is simple: is this infrastructure working harder for you this quarter than it was last quarter?
Without Momentum, the other four layers have a half-life. Built correctly once is not built correctly forever.
Find you → know you → understand you → trust you → compound.
Each layer is a prerequisite for the next. Platform without Recognition is an open door with no nameplate. Recognition without Intelligence is an entity with nothing to say. Intelligence without Signal is knowledge without authority. Signal without Momentum is a position that erodes the moment you stop tending it.
Want to know where your site breaks down? Start with Layer 1—free, instant, no login required.
Check your Platform layer →If you want someone to work through all five layers with you, that's the Infrastructure Audit.
Book the Infrastructure AuditCommon questions
Visibility to AI search engines requires fixing the infrastructure underneath your site—not the design or copy. The starting point is Platform: ensuring AI crawlers (GPTBot, ClaudeBot, PerplexityBot) can access and read your site, that your content is rendered in HTML rather than loaded by JavaScript, and that your sitemap and robots.txt are correctly configured. Once crawlers can get in, Recognition defines who you are to AI via schema markup and consistent entity signals. The PRISM Framework covers all five layers in sequence.
An AI search visibility framework is a structured approach to making a business findable, understandable, and recommendable by AI search engines—ChatGPT, Perplexity, Claude, and Google AI Mode. Unlike traditional SEO, which optimises for Google's ranking algorithm, AI visibility requires entity clarity, machine-readable content, and third-party validation across the broader web. A framework matters because the layers are sequential—fixing content before fixing crawler access, or building signal before establishing entity definition, produces work that doesn't compound. PRISM sequences the work in the right order.
SEO (Search Engine Optimisation) optimises for Google's ranking algorithm—links, keywords, page authority, click-through rate. AEO (Answer Engine Optimisation) optimises for AI citation: whether ChatGPT, Perplexity, and Claude can find, parse, attribute, and recommend a business when someone asks a relevant question. The technical requirements diverge significantly. AI crawlers don't execute JavaScript. Entity schema matters in ways it never did for traditional SEO. Off-site validation is evaluated differently. And the outcome is different—SEO produces rankings, AEO produces recommendations. PRISM is an AEO framework, though it builds on the same technical foundations that good SEO has always required.
PRISM is a sequential five-layer framework for making a founder-led business visible to AI search engines—ChatGPT, Perplexity, Claude, and Google AI Mode. The five layers are Platform, Recognition, Intelligence, Signal, and Momentum. Each layer enables the next. You cannot skip ahead.
Platform—the technical infrastructure that determines whether AI crawlers can access and read a site at all. This covers bot access via robots.txt, HTML-first rendering, sitemap.xml, llms.txt, and Core Web Vitals. Most small businesses are accidentally blocking the crawlers that matter. A site that looks polished to a human can be a blank page to ChatGPT.
Standard SEO optimises for Google's ranking algorithm—links, keywords, page speed. PRISM optimises for AI citation: whether ChatGPT, Perplexity, and Claude can find, understand, trust, and recommend a business. The technical requirements diverge significantly. PRISM also starts two steps before any technical work—with identity and positioning—because infrastructure built on the wrong identity has to be rebuilt.
Platform, Recognition, Intelligence, Signal, Momentum. Platform is technical AI access. Recognition is entity definition. Intelligence is machine-readable knowledge. Signal is third-party validation. Momentum is the compounding retainer layer that keeps the infrastructure current.
The PRISM Framework was developed by Aimee Q Devlin, a systems and infrastructure architect based in San Miguel de Allende, Mexico. It emerged from her work building AEO-first infrastructure for founder-led businesses across SaaS, publishing, education, and professional services.