How Perplexity searches
Perplexity is built as an answer engine: for almost every question it searches the web and adds numbered sources. This is how it works.
How Perplexity finds an answer
Always searching
Perplexity searches the web for nearly every question.
Choosing sources
From the results it picks the pages that answer the question best.
Writing the answer
The model writes a short answer from those pages.
Numbered sources
Each claim gets a number linking to its source.
Who visits your website
PerplexityBotAllowed on CogundiLets websites appear and be linked in Perplexity’s search results. According to Perplexity it is not used to train AI models.
Perplexity-UserAllowed on CogundiVisits a page when a user asks for it.
On Cogundi this is already taken care of
Every Cogundi profile lets the Perplexity crawlers in and is made to be read by AI: as a normal page, as Markdown, with structured data (JSON-LD) and through llms.txt. You do not have to set anything up.
How to become findable, and stay that way
A place AI may read
Your Cogundi profile is open to all AI crawlers. Even if your own website blocks them (for now), Perplexity finds your facts there.
The facts customers ask for
Cogundi shows which questions AI has about your business. You answer them: what you do, where, for whom and at what price.
One story everywhere
Your profile brings your details together with a source, so AI does not have to choose between conflicting versions.
Kept up to date, every month
After publishing we measure every month for a year what AI knows about you. New questions arrive in your feed, so your answers stay current.
Sources
Based on what the makers publish themselves. They do not fully disclose how an AI chooses its sources.