The Perplexity B2B buyer journey works fundamentally differently from ChatGPT or Google—and you can see it directly in the citation lists. Perplexity shows its sources immediately, making it crucial for B2B brands to understand how this AI search engine influences customers during consideration.
How Perplexity shapes the B2B buyer journey during awareness
| What Perplexity does | What that means for your buyer journey |
|---|---|
| Consults an average of 10 unique sources per query | The B2B buyer does not see your content; they see Perplexity’s synthesis of what it finds about you. |
| LinkedIn dominates the consideration stage (23 citations), not vendor websites | Visibility during consideration starts beyond your marketing automation. |
| Reddit ranks first in the decision stage (25 citations), ahead of all trade publications | Peer validation beats branded content when it matters most. |
“Two almost identical prompts can produce completely different citation lists, depending on one or two words in the question.”
— Finding from 90 Perplexity runs (April–May 2026)
Perplexity differs from the other two platforms I measure. With ChatGPT, Quolity data makes the retrieval of external sources visible behind the scenes. With Google AI Overviews, you can see citations within a summary, but the selection logic remains largely hidden. Perplexity presents its sources right in the interface: numbered, clickable and easy to check. The footnotes and links it selects are not random. That makes Perplexity the most transparent of the three and the most direct instrument for GEO analysis. You do not have to reconstruct what it did—it is there.
In the previous two articles in this series, I looked at how ChatGPT uses fan-out behaviour for B2B questions and how Google AI Overviews move through the B2B buyer journey. Now I am adding the third platform. Perplexity may not be dominant in the Netherlands, but it is still developing rapidly and could become more important.
How I collected the data
For this analysis, I asked Perplexity 90 questions: 30 prompts, each run three times. For more background on how AI search engines work, see my GEO guidebook. The prompts were divided across five clusters and two languages—20 Dutch and 10 English questions. The clusters were brand recognition (A), generic GEO definitions (B), comparison/expert questions (C), practical how-to questions (D) and Dutch expert prompts (E).
I manually exported the citations through the Perplexity interface. This is a smaller and more labour-intensive method than the automated captures for ChatGPT and Google AIO. The advantage is reliability: I know exactly what is in the data.
Perplexity cites an average of 9.8 unique domains per run. That is between Google AI Overviews (an average of 14.4 AIO citations per run in the earlier dataset) and ChatGPT, which is considerably more selective about what it picks up.
What does Perplexity retrieve during awareness?
During the awareness stage, I asked questions such as “What is Generative Engine Optimisation?” and “Which websites write about GEO and AI visibility?” These are questions someone asks at the start of their search, while they are still exploring.
Reddit tops the citation list, with 11 appearances in cluster B alone. It is followed by frankwatching.com, youvia.nl, conductor.com and Wikipedia. For definition questions, Perplexity does not retrieve only specialist sources, but a mix of community content, general marketing media and Wikipedia explanations. That is similar to Google AI Overviews during awareness, where 38% of AIO sources were encyclopaedic or explanatory content. Perplexity uses a broader mix: less vendor-driven, more community and trade media.
I noticed that frankwatching.com already has a strong presence during awareness—six citations for generic GEO questions. That signals its position in the Dutch GEO landscape: Perplexity treats it as a reliable source of basic information, not just as a trade publication for advanced readers.
What this means for a B2B brand: In the classic buyer journey, awareness is when a prospect first enters your marketing automation—through a blog, white paper or newsletter. Perplexity operates alongside that channel, but it does not cite what you publish yourself. It cites what trade media such as frankwatching.com and emerce.nl write about you. Visibility there is therefore as much a part of your strategy as the content feeding your marketing automation.
The Perplexity B2B buyer journey during consideration
The consideration stage includes comparison questions and expert prompts: “Which experts or agencies specialise in GEO?” “Who writes about GEO in Dutch?” These are the questions people ask when they know what they are looking for but not yet whom to choose.
The citation profile shifts clearly at this point. LinkedIn dominates cluster E (expert prompts), with 23 appearances across nl.linkedin.com and linkedin.com. Emerce.nl follows with 11 citations, then frankwatching.com (7), doublesmart.nl (8) and geoexperts.nl (7). It is mainly a combination of trade media and professional networks, with hardly any vendor content. This resembles Google AI Overviews during consideration, where consultancy.nl and frankwatching.com were the top AIO citations. Perplexity adds LinkedIn as a prominent source—something Google AIO showed less strongly.
One observation I want to share is the difference between two almost identical prompts. “Who are the GEO experts in the Netherlands?” produced a list in all three runs with Gijs Bodenstaff, Erik van der Veen and Peter van der Graaf as the main names, alongside Doublesmart and TEAM LEWIS as agencies. “Which experts or consultants advise B2B marketers on GEO and AI visibility?” produced a different picture in all three runs: Hans Schepers of Hands on GEO was the first expert mentioned, followed by Claudia Corpeleijn of Lead Today. I do not know whether my cookies or IP address gave me away. I am happy with the result either way 😉
The difference between those searches is the phrase “B2B marketers”. Perplexity interprets that detail as a different question and retrieves different sources. That is a useful insight for any B2B brand that wants to appear in expert prompts: the wording helps determine whether Perplexity retrieves your sources.
What this means for a B2B brand: In the buyer journey, consideration is when marketing automation works hardest—lead nurturing, email sequences and retargeting. At that same moment, Perplexity is not looking at your nurture campaign. It is looking at LinkedIn profiles that explicitly claim a B2B niche, trade publications and expert directories.
The Perplexity B2B buyer journey during decision
The decision stage includes how-to and direct problem questions: “How can I make our B2B brand visible in the AI tools our customers use?” “My website does not appear in AI answers; what can I do?”
Here Reddit again does what it did in the decision stage of Google AI Overviews: it takes a dominant position. In cluster D, Reddit has 25 citations—the most of any cluster. It is followed by frankwatching.com (8), aivisibilityhub.nl (6) and youvia.nl (4).
In Google AI Overviews, I saw Reddit grow from 60 appearances during awareness to 178 during decision. The absolute scale is smaller in Perplexity, but the logic is the same. Independent validation and practical implementation guides matter more now than vendor content and explanatory articles.
I also saw aivisibilityhub.nl appear six times for how-to questions about AI visibility—a relatively new domain that I did not see as prominently on the other platforms. It is the type of site that performs well for specific implementation questions in Perplexity, even though it has not reached the top in Google AIO.
How does citation style change with language?
One technical difference is worth mentioning: Dutch Perplexity exports use footnotes as citations, while English runs use inline Markdown links. That has content implications.
English answers to prompts such as “What is Hands on GEO?” are considerably richer and more structured than the Dutch equivalents. Perplexity generates a table of features, cites several pages from the site and gives a concrete description of the platform, including the founder’s name and background. I do not see the same level of detail in Dutch runs.
My assessment is that this is partly due to the amount of English-language content Perplexity draws on: more context, more mentions and richer training data. This is an argument for maintaining English-language content consistently alongside Dutch, even if your primary market is Dutch-speaking.
What does this mean for your B2B strategy?
The common thread I see in the other two articles is this: the further a prospect moves through the buyer journey, the less a B2B brand can rely on its own content. External validation makes the difference. I saw this pattern in ChatGPT too—see my analysis of how ChatGPT shapes the B2B buyer journey. In Perplexity, the pattern is visible directly in the citation list rather than only after interpretation.
Perplexity adds a nuance I saw less clearly on the other two platforms: the exact wording of a question noticeably changes the retrieval path. Two almost identical prompts can lead to completely different citation lists because of one or two words.
That makes content optimisation for Perplexity less about “more content” and more about the right connection. Are you visible in a B2B context, in the places where Perplexity looks for confirmation of that context? Then you appear. If not, you simply are not there.
Frequently asked questions
Why does Perplexity show its sources when ChatGPT does not?
It is a deliberate product distinction. Perplexity positions itself as an “answer engine with citations”, putting transparency at the centre. ChatGPT shows sources only if you explicitly choose browsing—and even then, its internal search queries are not directly visible.
How many sources does Perplexity cite per run on average?
In my dataset of 90 runs, an average of 9.8 unique domains per run. This varies by cluster: expert prompts drew an average of 11.8 sources per run, compared with 8.1 for comparison questions.
Is Reddit as dominant in Perplexity as it is in Google?
Reddit ranks first overall in my dataset of GEO-related questions, with 59 citations across all clusters. It is strongest for how-to questions in the decision stage, with 25 citations in cluster D. This is comparable to Google AI Overviews, where Reddit was cited almost three times as often during decision as during awareness.
What changes when the question is in English?
English answers are considerably longer and more structured than Dutch ones. For English brand questions, Perplexity even generates tables of features, cites multiple pages from a website and adds more background. My explanation is that the English-language content landscape is simply richer and gives Perplexity more material to work with.
Does the wording of a question matter?
Yes—and this may be the most useful observation from this measurement. Two almost identical prompts about GEO experts in the Netherlands produced completely different citation lists, purely because of the words “B2B marketers”. Perplexity treats that detail as a different question and retrieves different sources as a result.
Is Perplexity a useful platform to optimise for?
Perplexity has a limited market share in the Netherlands compared with Google and ChatGPT. But as a diagnostic tool for seeing which sources an AI platform chooses for questions in your market, it is the most transparent of the three and gives B2B brands a useful indication of their online authority. Actively steering visibility in Perplexity, however, remains difficult.
Sources
- Own dataset: 90 manual Perplexity runs (p01–p30, runs 1–3), conducted April–May 2026.
- Method: URL extraction from footnote citations (Dutch runs) and inline Markdown links (English runs).
- Comparative ChatGPT data: The AI Buyer Journey: How ChatGPT Navigates Your B2B Brand.
- Comparative Google AIO data: Google AI Overviews and the B2B Buyer Journey.
By Hans Schepers—a digital marketing professional specialising in GEO (Generative Engine Optimisation) and AI visibility for B2B brands.
This translation was made by AI.