How does Perplexity work and how do you get into the sources?
Perplexity shows its sources with every answer and rewards authority, timeliness and structure. What that concretely asks of your pages.
Lees dit in het Nederlands →$ trace --answer=sources
How does Perplexity work and how do you get into the sources?
Perplexity answers every question with a list of sources attached, and chooses those sources on authority, timeliness and structure. Getting into the sources therefore asks three things: content that is correct and rests on something, a page that is machine-readable, and a date that shows the piece is current. This piece explains the mechanism; whether Perplexity matters for your market is something you decide for yourself at the end.
Of all AI search engines, Perplexity is the most transparent. Where other assistants sometimes show their sources and sometimes don't, Perplexity puts them at the top of every answer, numbered and clickable. That makes the platform interesting for anyone working on their visibility: you see in black and white whether you're there.
01 / the mechanismEvery sentence rests on a source
Perplexity is built as an answer engine on live web data. For every question, the system searches the web live, selects a handful of pages, and assembles an answer from them in which every claim links back to a numbered source. The user can check each claim directly.
That differs fundamentally from how ChatGPT works by default, where a large part of the answers comes from training data and sources only appear once web search kicks in. How that selection works there is set out in how does ChatGPT arrive at its answers. For Perplexity the simple rule applies: no source, no place in the answer.
02 / the selectionWhat Perplexity rewards
Perplexity gives priority to sources that are authoritative, recent and well structured. Those three words sound abstract, but they can be made concrete:
| What Perplexity rewards | What that concretely asks of your page |
|---|---|
| Authoritative | The page rests on something: a named author, references to traceable sources, and content that goes beyond what's already everywhere |
| Recent | A visible date that isn't years old; between two comparable pages, the current one wins |
| Well structured | The answer is readable as a standalone fragment too: headings that name the question, paragraphs that open with the conclusion, tables where there's something to compare |
Perplexity shows with every answer who the sources are. Your visibility there isn't a feeling but an observation.
03 / in practiceWhat that asks of your pages
Anyone who wants to get into the sources works on three sides of the same page.
The content: answer the question the reader asks, fully and in the first paragraph. Back up what you claim and add a named author. The form: use headings that literally contain the question, keep paragraphs to one thought, and show a publication date or a last-edited date. The technique: make sure PerplexityBot is allowed to read your site in robots.txt and that your pages load without detours, so not hidden behind scripts or a login.
What doesn't work: stuffing the same page with search terms or creating dozens of thin pages. Perplexity selects only a handful of sources per question, and thin material doesn't make that selection.
04 / the testWhy Perplexity is your most honest mirror
Because every source is visible, Perplexity is the fastest place to test your own position. The test in five steps:
- Formulate the five questions your client actually asks, in their own words.
- Ask each question to Perplexity and open the source list above the answer.
- Note per question who is in the sources, and whether you're among them.
- If you're not there: click open the sources and compare their pages with yours on the points from the table above.
- Repeat the round monthly, with the same questions and the same method.
That observation is a snapshot, not a measurement. Answers differ by phrasing and by day. If you want to turn it into a figure you can compare over months, fix a set list of questions in advance and repeat it with the same method. How to set that up is set out in how do you measure AI visibility.
05 / the limitWhen this isn't your priority
Perplexity has a smaller audience in the Netherlands than ChatGPT and Google. If your market is mostly people who never search with AI assistants, this platform isn't a priority and your regular visibility will deliver more. Anyone whose basics aren't in order yet (a site that loads slowly, pages without dates or authors) is also better off starting there; the same improvements then count on every platform.
The reason to take Perplexity seriously anyway isn't today's scale, but what the platform makes visible: the same qualities that Perplexity rewards also weigh with the other AI systems. Whoever writes for the strictest judge is usually better off with the rest too.
Does Perplexity always show sources?+
Do I need to optimise separately for Perplexity?+
Why am I in Google but not in Perplexity?+
How do I test whether Perplexity uses me as a source?+
Does the date on my article really matter?+
Is your company in the sources when your client asks their question to Perplexity?
The scan
This is the baseline measurement: fixed search queries on ChatGPT, Gemini, Claude, Perplexity and Google. Current prices are on the pricing page. Even without a follow-up, you keep the report.
This text was produced with AI assistance and checked and approved by a human before publication.