Why QueryBurst Does Not Track AI Visibility
QueryBurst is not an AI visibility tracking platform. It does not continuously monitor a predefined panel of prompts, conversations, brands, citations or recommendations and turn those observations into a longitudinal measure of AI visibility.
QueryBurst is built for investigation rather than tracking. Users define the customer situations and journeys they want to examine. QueryBurst then models the answer process across those journeys so they can inspect where a brand enters, survives or disappears, what information is retrieved, what evidence is selected, and what changes may alter the outcome.
That distinction matters because tracking and investigation answer different questions.
Tracking asks: what happened over time?
QueryBurst asks: what appears to be producing this outcome, and what can I change?
QueryBurst does not replace prompt tracking with another kind of tracking
QueryBurst rejects prompt tracking, but does not replace it with conversation tracking, journey tracking, citation tracking or some other variation of the same measurement model.
Our position is broader.
A prompt panel is one way of repeatedly observing outputs under defined conditions.
QueryBurst instead lets users construct a customer journey and investigate how the answer process develops across it.
A QueryBurst run is therefore closer to an experiment than a monitoring panel.
It is designed to test a situation, inspect the process, compare explanations and identify where an intervention may matter.
QueryBurst does not track conversations
Answer Engine runs stateful, multi-turn conversations.
That does not mean QueryBurst tracks conversations.
The conversation is part of the experiment.
The user defines the customer, the situation and the journey being investigated. Each turn inherits what happened before it, including previous assistant responses, products that entered or left consideration, objections raised, evidence introduced and branches taken.
QueryBurst models what happens through that evolving state.
It does not claim that the resulting journey is a measurement of how a wider population of users behaves.
It is a modelled research journey used to investigate a defined problem.
QueryBurst does not track customer journeys
Users model customer journeys in QueryBurst.
QueryBurst does not discover or monitor the real journeys of a customer population.
A buyer journey in Answer Engine is a deliberately constructed, stateful path used to investigate how different contexts and decisions may influence an AI-mediated research process.
That journey can branch.
A user can test what happens when a requirement changes, when the assistant introduces a new product, when an objection appears, or when a different path is taken from the same earlier state.
Those branches are experiments.
They are not claims that QueryBurst has observed the population distribution of real customer behaviour.
QueryBurst does not track citations
QueryBurst may show citations, sources and evidence produced within an Answer Engine investigation.
It does not continuously count citation frequency across a predefined panel and interpret that count as a measure of overall AI visibility.
A citation is one observable event inside a wider answer process.
A source can be retrieved without being cited.
It can be selected without being recommended.
It can support a factual claim while another brand wins the commercial decision.
QueryBurst therefore treats citations as evidence to investigate rather than a longitudinal visibility metric to track.
QueryBurst does not track AI visibility over time
QueryBurst analyses can be rerun.
Two runs can be compared.
A user may repeat an investigation after changing a page, adding evidence, strengthening external corroboration or testing a different customer journey.
That is not the same thing as continuously tracking a fixed panel and interpreting changes in the resulting time series as changes in a brand's overall AI visibility.
QueryBurst's position is that rerunning a controlled investigation can be useful when the question is:
That is different from asking:
QueryBurst is designed for the first question.
A QueryBurst run is an experiment, not a measurement panel
Tracking is useful when the objective is to observe the same defined thing repeatedly and see how the observation changes over time.
Investigation is useful when the objective is to understand a mechanism, test competing explanations, identify where something failed, and determine what may change the outcome.
QueryBurst chose the second model. Our definition of AI visibility treats it as conditional on the customer, the conversation and the information environment — which means investigation into specific conditions is more useful than tracking a fixed panel over time.
That is why it exposes the answer process rather than reducing repeated outputs into a visibility chart.
Each Answer Engine run begins with a defined research situation.
The user models the customer and the journey.
QueryBurst models what happens next.
That can include the model’s pre-retrieval position, accumulated conversation state, fan-out and discovery, retrieval, source selection, evidence use, grounding and the final answer — all inspectable within the run.
The purpose is to make those stages inspectable within the QueryBurst run so that a user can test where a brand enters or drops out and what may be causing it.
The purpose is not to claim that the run represents the market as a whole.
Why QueryBurst makes this distinction explicit
AI visibility products are commonly understood as tracking products.
That means when QueryBurst says it does not use prompt tracking, it is easy to assume that it must instead track some other object:
- conversations
- journeys
- citations
- retrieval events
- recommendations
- AI visibility itself
It does not.
Those things may appear inside an investigation.
They are not what QueryBurst continuously measures as a market-level time series.
QueryBurst's position is that visibility should be investigated at the level where it is gained, lost or changed.
That requires a different methodology from tracking outputs.
QueryBurst's position
QueryBurst does not track AI visibility.
It investigates it.
Users define the customer and the journey. QueryBurst models the process that produces each answer. That can be as simple or as deep as the question requires.
The result is not a longitudinal visibility metric.
It is diagnostic evidence about what appears to be happening, where the brand enters or disappears, and what can be tested next.
Tracking records the output. QueryBurst investigates the process.