The most discerning buyers build their own collections of trusted sources on Perplexity — ten, fifteen hand-picked references on which they base their purchasing decisions. If your brand isn't in those collections, you don't exist for them at the moment that matters most. This isn't passive visibility: whoever ends up in those lists gets consulted every time the buyer has a question. Getting in is possible, and it doesn't require an advertising budget.
Perplexity Collections: thematic clusters of sources curated by users and experts. If an expert in your sector creates a collection and you’re not in it, you’ve lost 50 qualified queries a month for several months.
Think about what that means in concrete terms. You’re a high-end footwear manufacturer in the Vigevano district, in the province of Pavia, the historic heart of Lombardy’s footwear sector. An American product sourcing consultant opens Perplexity, creates a Collection called “Italian luxury footwear manufacturers — handcrafted leather”, and adds ten curated sources to it — company websites, trade articles, technical sheets. From that moment on, every time he or one of his colleagues queries that Collection, the AI answers draw only from the sources he put inside. If your company isn’t there, you don’t exist for that buyer. Full stop.
In this article I’ll explain what Collections are (the historical name, now renamed Spaces in the current interface), why they’re one of the few AI visibility channels you can literally own instead of hoping to earn, and what you can do today to avoid being left out of the Spaces in your sector.
What Spaces are and why they count as a marketing asset
A Perplexity Space (or Collection) is a thematic workspace where a user gathers a set of manually curated sources — links to papers, articles, websites, uploaded PDFs. When you query that Space, Perplexity answers primarily using those sources as grounding, not the entire open web.
Translated into practice: whoever creates a Space builds a vertical mini search engine on their topic, where the list of admitted sources was decided by them. If you’re a B2B supplier, a consultant, a niche manufacturer, this is one of the first times you can own an AI recommendation channel instead of chasing it.
It’s not an academic paper telling you this, but the principle is a direct deduction from how Perplexity’s RAG (Retrieval Augmented Generation) systems work. In the previous articles of this series I explained how Perplexity builds answers starting from a selection of sources retrieved in real time: if the source pool is limited to those in the Space, the output is limited to those sources. From this it follows that being inside the Collections of your sector is equivalent to being inside the index of the vertical search engine that your potential customer will use every day.
Why an expert creates a Collection (and why you need to get in)
Collections come about for a very simple reason: people who work in a technical sector are tired of receiving generic AI answers full of weak sources. A fashion consultant, a buyer, a vertical journalist, a professor — all these profiles create Spaces to work on a set of reliable sources they have pre-selected.
Once the Space is created, two interesting things happen for you:
- The expert can share it publicly with their own audience (newsletter, LinkedIn, professional communities). Every user who opens it generates queries with your sources as the basis, if you’re there.
- The expert themselves queries that Space dozens of times a month for their daily work. If you’re in, your company gets cited systematically in their answers.
This ties into a mechanism I described in the article on implicit reference weight: the frequency with which a domain is chosen as a source in a curated context weighs heavily on its authority as perceived by the model.
Creating a Space and not sharing it.
The case of the Vigevano footwear manufacturer that got into a Collection
Let me tell you about a concrete case, anonymized for confidentiality. A footwear manufacturer from the Vigevano district, high-end — Goodyear welted shoes, Italian leathers, production of 8,000 pairs/year — sells to European multi-brand boutiques and to private label for French fashion houses. Typical customer: a buyer in Paris or Monaco who needs to find 3-4 Italian partners for small-series production runs.
The starting problem was a classic one: excellent product, decent website, zero presence in AI answers when buyers asked “Italian small-batch shoe manufacturers Goodyear welt”. Perplexity always cited the usual three big names, never them.
Here’s what happened. A French journalist who writes about Italian craftsmanship for a trade magazine created a public Collection titled “Italian artisan shoemakers — small-batch production”. He put 14 sources inside: some company websites, two Il Sole 24 Ore articles on the Vigevano district, a Wikipedia entry on the Goodyear welt, two PDFs from chambers of commerce. The footwear manufacturer in question was inside: the journalist knew it from a factory visit two years earlier, and had found their “Production Capabilities” page detailed enough to merit inclusion.
Result observed over the following three months: an increase in qualified inquiries via email from the French and Swiss markets, all with the phrase “we found you through a Perplexity research” or equivalent. Not an explosion of traffic, not a change of scale — but already pre-qualified inquiries, with buyers who arrived knowing what the company did and what it didn’t.
The operational consequence is clear: it’s not the volume that changes, it’s the quality of the lead. A buyer who contacts you after reading your profile inside a Collection curated by a sector expert arrives with 70-80% of the qualification work already done.
You need to put 15-20 real sources from the sector in it (competitors included, specialized press articles, papers, institutional data).
The test you can run in 15 minutes
I want to be honest: there is currently no public way to search all of Perplexity’s Collections about your sector — there’s no navigable index like Google’s. But you can do two useful checks.
Test 1 — Check whether you’re in the known public Collections. Open Perplexity, search for your sector + “collection” or “space” on the public channels where experts share their Spaces (LinkedIn, Twitter/X, sector newsletters, vertical Discord communities). For your sector (artisan footwear, in this case) find out who the 5-10 most active journalists, consultants, and buyers are. Check whether they share Spaces.
Test 2 — Create your own Space and use it as a benchmark. Go to Perplexity, log in with a Pro account (the Space requires a paid account), create a new Space called “Italian artisan footwear — Goodyear welt production” or the equivalent for your sector. Add 15-20 sources that you think should be in an authoritative Collection on the topic. Query it with the questions a potential customer of yours would ask: “best Italian Goodyear welt manufacturers small batch”, “Italian artisan shoemakers Veneto Lombardia”, “where to source handcrafted leather shoes Italy”. See whether your company comes up. If it doesn’t come up even inside a Space where YOU put the sources, you have a problem with your company content (your pages don’t give enough signal) and not just a channel problem.
This is an entry-level test — the real analysis of your positioning in AI search requires professional tools and a more structured audit.
The observation I make about manufacturing clients
Over the last 6 months I’ve seen a recurring pattern in niche B2B clients (manufacturing, specialty food, professional services): those who created 2-3 Spaces on their own topics and shared them with their qualified audience saw an increase in “AI-mediated” inquiries — customers who arrive explicitly stating that they discovered them via Perplexity, ChatGPT with browse, or AI aggregators.
Limits of the observation: a sample of 12-15 clients, very diverse sectors, it’s not a controlled study. It’s a pattern, not statistical proof. But the mechanism is consistent with how Perplexity’s RAG systems work, so I feel fairly confident recommending it as a low-cost, high-potential action.
The mistakes I see most often
- Creating a Space and not sharing it. A private Space is useful for internal work, but it generates no visibility. You have to share it publicly with your audience to activate the virtuous circle.
- Putting only your own content inside the Space. A Collection where you’re the only source is useless and suspicious. You need to put 15-20 real sources from the sector in it (competitors included, specialized press articles, papers, institutional data). You are one of the sources, not the only one.
- Thinking of the Space as a company “press kit”. It isn’t. It’s a thematic research tool. If it doesn’t offer real value to those who use it to work, no one will use it.
- Ignoring other people’s Collections. If a journalist in your sector has already created a Collection on the topic and you’re not in it, the first step is to understand why — usually it’s not malice, it’s that your site didn’t have a “Capabilities” page clear enough to merit inclusion.
What to do concretely this week
- Identify the 5-10 most active experts in your sector who use Perplexity (journalists, consultants, buyers, professors). Follow their public channels.
- Create 2-3 Spaces on your main topics. Add 15-20 curated sources to each (including authoritative third-party sources, not just your own).
- Share the Spaces with your qualified audience via newsletter, LinkedIn, sector events.
- Review your main pages to make sure they are “Collection-ready”: a clear capabilities sheet, verifiable production data, the presence of entities recognizable by the model (see the article on Named Entity Recognition).
- Compare your “About” and “Production” pages with those of the 3-5 competitors you see cited in the Collections in your sector.
Why this brings you closer to AI answers
Perplexity’s Collections are one of the rare cases in which visibility in AI answers passes through a channel you can truly own and influence directly. You don’t depend on the algorithm, you don’t depend on the domain’s general authority: you depend on the fact that your content is solid enough to merit a place in a Collection curated by an expert.
In the next articles of this series I’ll explain how Perplexity Pages works, how Perplexity chooses between different Focus Modes depending on the query, and how the ChatGPT answer architecture differs from Perplexity’s when it comes to citing sources.
