QueryBurst
The Workspace for Improving Brand Visibility in AI Search
Start new thread

QueryBurst - The Workspace for Improving Brand Visibility in AI Search

by•
QueryBurst is a workspace for modelling multi-turn customer journeys through AI search to discover where your brand enters, influences, and leaves the conversation — and why. Create target personas and branching conversations, explore how context changes answers, and inspect the research process from query fan-out to source selection and citations. It's like stepping inside your customer's AI conversations to see what shapes the answers they receive. Think people and conversations, not prompts.

Add a comment

Replies

Best
Maker
📌
I'm David McSweeney, founder of QueryBurst. I’ve been involved in search for almost 30 years, and was previously head of content for Ahrefs and Seobility. QueryBurst first launched here in July 2025 as a Google Search Console analytics tool. But I got increasingly exasperated with the host of VC funded tools for “AI visibility” (read prompt trackers) and started building out the tools actually needed to do the job without AI slopifying your website, spamming the internet, and compressing everything into a meaningless “AI visibility score”. The first stage was Site Intelligence and the AI Search Toolkit. A suite of tools built around analyzing how AI actually sees your website (and working on improving it all): from a knowledge graph of the entities and claims your site makes, to tools for retrieval optimization, reducing hallucinations, optimizing for AI agents, query fan-out, and more. Basically, all the tools everyone else was standing up on stages and talking about hypothetically, I was actually shipping them. But there was still a problem: people saw “AI visibility” and assumed QueryBurst was a prompt tracker. To clarify it wasn’t, isn’t, and never has been. I’ve been vocal about how prompt tracking is a flawed model. So a few months ago, I turned my attention to a different problem: how do we actually investigate a brand's visibility in AI search without reducing everything to isolated prompts and arbitrary scores? And in the end, it was all rather obvious: People are having conversations with AI - so the unit of analysis should be a conversation, not an isolated prompt. Who's asking matters. So we should model different customer personas. LLMs already have knowledge and associations about brands before searching the web. So we should investigate those priors before introducing external evidence. Context should carry through the conversation. Conversations should be multi-turn and branching (think buyer journeys). And none of this should be compressed into a “visibility score” or be tracked obsessively every day. The result was QueryBurst Answer Engine, which lets you create customer personas, model multi-turn conversations, branch into different research paths, and inspect how answers develop, including query fan-outs, retrieved evidence, source selection, and citations. Once you've set up several personas and conversations, the data starts to give you a rich, actionable view of how you stand in AI search, including which websites are shaping the answers. Everything is inspectable. But that doesn’t mean it’s hard to use. The simplest thing you can do in QueryBurst is ask a question the way your customer would and read what comes back. Everything else — personas, branching conversations, retrieval inspection — is there when you want it. But none of it is required before you start. Your customers are having conversations with AI. QueryBurst lets you step inside them.