How to Track Your Brand Ranking Across Every Major AI Search Engine and Close the Gaps

A founder Will Leatherman spoke with recently searched his payments company on ChatGPT. It came back at number 3. He felt great. Then Will had him open Perplexity and run the same search. Five different companies appeared. His brand was not one of them.

He had been treating AI search as a single scoreboard when it is 4 separate ones, each with its own memory and its own shortlist for every buyer query.

Why does your brand rank differently on every AI platform?

Each LLM draws on a different training corpus, retrieves from different external sources, and weights relevance signals differently when generating category recommendations. Run "what is the best CRM for modern B2B companies" on Perplexity and you might see HubSpot, Salesforce, Pipedrive, Zoho, and Microsoft Dynamics. Run the same query on ChatGPT and Attio might top the list while Zoho disappears.

Neither result is wrong. They are different models answering from different contexts.

The practical implication is direct. Catalyst data from 150 B2B software clients ranging from Series A to Series D shows that "4 out of 10 B2B buyers now start inside of an AI tool" before making a purchasing decision. If your brand is invisible on even one platform, you are missing a share of those buying conversations before they begin.

Checking one platform once a month is not a visibility strategy. Running the real buyer queries across every major LLM on a repeatable cadence and building a single ranked view of where you stand is.

What does AEO actually measure?

AEO (answer engine optimization) is the practice of earning placement in AI-generated answers, not search result links. When a buyer types a category question into an LLM, the engine returns a ranked shortlist. AEO is the discipline of appearing on that shortlist consistently across platforms.

Three factors determine your score:

Technical. LLMs use bots to crawl and index your website before they can recommend you. Missing structured data, a blocked crawler, or the absence of an llms.txt file means the engine has no information about your brand and will not cite you regardless of how strong your content is.

On-site content. Original content published on your domain gives LLMs something to reference. As Will put it in the session, "AI is quite lazy and it will just copy and paste. It will just find the most credible answer and reference that when people search." If an engine can generate an answer from memory, it will not go looking for a citation. Proprietary data, founder perspective, and unique methodology force it to reference your site.

Off-site mentions. Third-party publications and niche industry sources carry significant weight in how LLMs rank category leaders. A global payroll client in Catalyst's portfolio was competing against large legacy providers who had dominated the category for years. After securing placements in a handful of niche publications, they surpassed several of those incumbents, including Rippling, in AI-generated shortlists for their category.

The goal is not ranking number 1. The goal is appearing in the top 5, because that is the shortlist buyers use to decide which vendors to evaluate.

How do you build a weekly AI ranking agent?

The agent built in this workshop uses Claude's computer use capability to control a browser, run buyer queries on each platform, collect the ranked responses, and compile a blended leaderboard. The build runs inside Claude Cowork with a shared starter project sent to all attendees after each session.

Step 1. Generate the buyer queries

The agent starts by identifying the real questions your prospects type into AI tools before building a vendor shortlist. For a CRM, those look like:

• "Best CRM for B2B SaaS companies"
• "CRM with easiest implementation"
• "Alternatives to Salesforce for fast-growing startups"
• "What CRM works best for companies under 100 employees"

The specificity matters. LLM search queries are conversational and pain-point specific. "Best banking as a service consultancy for decoupling legacy systems" is a real high-volume query for that category, and it is a query that a niche provider can win even if a generic keyword was dominated by legacy players for years.

Step 2. Run each query across all 4 platforms in fresh sessions

The agent opens ChatGPT, Perplexity, Gemini, and Claude in new sessions with no memory carryover and submits each buyer query. Fresh sessions matter because LLMs with memory can return results biased by prior conversation history, which gives you an inaccurate picture of what a first-time buyer actually sees.

Step 3. Build a blended leaderboard

The agent collects the ranked results from each platform and merges them into one view. You can see where you appear, where you are absent, and which competitors hold the positions you want.

Step 4. Schedule the agent to run weekly

The first run takes longer because the agent is building its query set and establishing baseline results. Once configured, the agent runs on a set cadence. Each week you get an updated leaderboard showing whether your position improved, held, or dropped after any content or technical changes you made in the prior week.

What should small brands do if they are not showing up at all?

The concern that only the top 5 or 6 brands ever appear in AI shortlists is understandable. SEO worked that way. AEO does not.

LLM queries are longer, more specific, and more varied than keyword searches. A smaller brand can rank for "best animated health education platform for elementary school" even if it could never rank on Google for the generic head term. The query space is far larger, and that creates room for niche players.

One live audience participant in this session ran a kids' health education platform. The AEO audit showed that brand was already ranking 24% of the time across Claude, Google, and Perplexity with no dedicated on-site or off-site content effort at all. A focused content plan targeting the specific queries where they already appeared would compound that baseline quickly.

Here is the priority order for a brand that is not yet showing up:

| Priority | Action | Why it matters | |---|---|---| | Fix technical first | Add llms.txt, FAQ schema, structured data, article schema | LLMs cannot cite what their bots cannot read | | Publish original on-site content | Write answers to the buyer queries your audit identifies | AI cites answers it cannot already generate from memory | | Secure off-site mentions | Get listed in niche publications and best-of roundups | Third-party citations carry heavy weight in LLM ranking decisions |

The sequence matters. Technical gaps block everything downstream. A brand with strong content but a broken crawler still shows up nowhere.

For a full breakdown of which content types drive the most AI citations, see The 4 Types of Pages That Drive AI Citations.

The takeaway

Each major AI platform keeps its own ranked shortlist for every buyer category. Checking one is not enough. The agent built in this workshop runs your brand's real buyer queries across all 4 major LLMs on a weekly cadence and gives you one consolidated view of where you stand and where you are invisible.

Start by running your site through Catalyst's free AEO audit tool to get your baseline score across technical, on-site, and off-site signals. That audit also generates the buyer queries your prospects are actually using and builds a content plan tied to each gap. From there, build the weekly tracking agent so you can see your position move as you implement fixes.

For a deeper look at the mechanics of ranking in AI answers, How to Get ChatGPT, Claude, and Gemini to Recommend Your B2B Company and How to Audit Your AI Search Ranking in 20 Minutes are the natural next reads.