What LLM Search is
AI assistants like ChatGPT, Perplexity, Google AI Overview, Google AI Mode, Gemini, and Claude are quickly becoming where people research before they buy. Instead of scanning a page of links, they ask a question and take the answer. That raises a new question for your brand: when someone asks AI about your category, do you show up, and how do you compare to your competitors?
LLM Search is the Slate report that answers exactly that. Slate tracks all six of the major AI platforms for you: Claude, Gemini, ChatGPT, Perplexity, Google AI Overview, and Google AI Mode. You give it the questions your buyers ask AI, Slate runs those questions across every one of those engines on a schedule, records each answer, and turns thousands of responses into a handful of clear metrics you can act on.
What it does
At its core, LLM Search continuously checks how AI answers treat your brand and packages the results into six views:
- Overview: your scoreboard, built on four headline metrics. Brand Mentions shows how often answers name you, Visibility Score shows how often you appear at all, Where you rank in AI shows your position versus competitors, and Share of Voice shows your slice of the whole conversation. It also trends those metrics over time, breaks them down by engine, and shows who leads each topic.
- Citations: the sources AI pulls from, sorted into your pages, competitor pages, social, and other, plus which pages mention you and which do not.
- Fanout Queries: the hidden sub-searches engines run behind the scenes when answering, so you can see the full set of questions being researched.
- Sentiment: whether you are described positively, neutrally, or negatively, and around which themes.
- Prompts Tracked: every question you are monitoring, with the engines it runs on, its schedule, and its mention and citation counts.
- Answer History: the raw, timestamped AI answers themselves, so you can read exactly what was said and see why a number moved.
Every view can be sliced by platform, topic, prompt, location, and time period, with a built-in compare to the previous period.
How you use it
1. Add the prompts you want to track. Click Run Prompt and either add a single prompt (typing the question the way a real buyer would ask it, then choosing the engines, a topic, a schedule, and a location) or use Bulk Upload (CSV) to load a whole list at once. Focus on decision-stage questions such as "best," "top," "alternatives to," and "how to choose," since those are the answers that actually influence buyers.
2. Read the scoreboard. Start on Overview to see whether your visibility is up or down, where you rank against competitors, and which engines you are strong or weak in.
3. Diagnose. When something moves, use Citations to find the sources AI trusts (and the prompts where competitors are cited but you are not), Fanout Queries as a content checklist of sub-questions to answer, Sentiment to catch how you are being described, and Answer History to read the actual responses behind the numbers.
4. Act. Turn the gaps into work: content to create, pages to fix, and sources to earn a mention from. Then watch the metrics respond on the next run.
Getting started
A quick way to get value in your first session:
- Add five to ten prompts covering your most important buyer questions, grouped under one or two topics, and set them to run on a recurring schedule.
- Let them run once (or run them now), then open Overview to establish your baseline: mentions, visibility, rank, and share of voice.
- Open Citations, filter to prompts where you are not mentioned, and note one or two content gaps to close first.
- Skim Fanout Queries for sub-questions your content does not yet answer.
- Check back weekly: read Overview for movement, then use Citations and Answer History to understand the why.
Summary (quick recap)
In short, LLM Search tracks how your brand shows up across Claude, Gemini, ChatGPT, Perplexity, Google AI Overview, and Google AI Mode, then turns those answers into action:
The workflow: add your prompts, read the Overview, diagnose with the other tabs, then turn the gaps into your next piece of work.