Falkview.
Menu

One method, run every week

Falkview Group tracks both channels every week: your position in search for the questions that matter, and what ChatGPT, Perplexity, Gemini, Copilot and Claude answer when a prospect asks the same thing directly. We keep every result, and the accumulated record becomes a tracked position: whether you appear, in what terms, next to whom, and on the strength of which sources.

The method

The Vantage Method

Most work in this field starts with an opinion about what to change. The Vantage Method starts with a measurement, and returns to one every week, so every decision is read against a record instead of a hunch.

BASELINEMeasured before anythingchanges01PROMPT BANKWritten for this client02STRUCTURERead correctly by both03LOOPRuns again, every week04
Baseline
We measure both channels before anything changes.
Prompt design
We write questions for this client specifically.
Structure
We engineer the presence so a crawler and a model both read it correctly.
Loop
Weekly runs compare citation sources across both channels, we act on the gaps, and the cycle returns to Baseline.

One pipeline, two channels, run every week.

The Loop stage runs on this pipeline.

PROMPT BANKQuestions written foryou01WEEKLY RUNSPut to search andassistants02ANALYSISShare of voice,sentiment, sources03DASHBOARDA live position, nota report04

Four layers, one system.

Each stage feeds the next, from a prompt bank written specifically for you through to a live dashboard you can check on any day.

Layer 01

Prompt bank

A custom bank of questions, written specifically for your engagement: the ones a prospect would actually type into a search bar, or ask an assistant, before deciding.

Generic keyword lists do not describe how someone actually searches or asks, and neither does a question set shared across clients. Each engagement gets its own bank, written from scratch for that client, tuned for a search query as well as a spoken-language assistant prompt, and revised as the market's language shifts.

Layer 02

Scheduled runs

We put your questions to search and the assistants live, every week, and keep every result.

Both channels drift: rankings move, and model answers change as providers update them. A single snapshot tells you nothing about direction, so we query the same bank live — not from a pre-aggregated database — on a fixed weekly cadence across search and ChatGPT, Perplexity, Gemini, Copilot and Claude, and we retain every raw result.

Layer 03

Analysis

Raw results from both channels become share of voice, sentiment, and cited sources, the most actionable of the three.

We examine each search result and assistant answer for whether you appear, in what terms, alongside whom, and on the strength of which sources, then aggregate the pattern into a trend rather than an anecdote. Share of voice and sentiment describe where you stand; citation sources describe what to do about it. If an assistant keeps citing a directory you are not listed in, that is a listing to go get. If it is leaning on a competitor's page for a claim you could make more clearly yourself, that is a specific paragraph to write. It is the one line in the report that turns directly into a task, rather than a number to keep watching.

Layer 04

Dashboard

A live view of where you stand in both, not a monthly PDF.

You can log in on any day and see the current position across search and assistants, and how it moved. Most monitoring in this space arrives as a document once a month; this updates continuously as each week's run lands, so the record builds in the open rather than sitting in a file you wait for. That is what makes month seven more useful than month one.

Questions

Why query the models directly instead of using an existing AI visibility tool?
Most AI visibility tools read from large, pre-aggregated mention databases. Those databases are built around consumer brands with real search volume, and an individual professional or small firm has close to no footprint in them. Querying the assistants live, with questions written for that specific client, is the only way to see what is actually said about someone the aggregate data has never heard of. It runs alongside the standard, more direct check of where the same client stands in search.
How often is it measured, and why weekly?
Every week, across both channels. Search positions shift and model answers drift, as providers update models and as the sources behind an answer change, so a single measurement describes a moment, not a direction. A weekly cadence is frequent enough to catch a change while you can still act on it, and steady enough that the trend line means something.
What is actually reported?
You either appear or you don't, in the search results and in the assistant's answer alike. Beyond that: the terms used to describe you, who appears alongside you, and which sources the assistant leaned on. Together, those are what you can improve; a position number on its own is not.

Find out what the answers already say about you.

Get your free report