Perplexity Discover Feed: How to Get Your Brand Featured

By UpGeo · 2026-07-16

What It Takes to Appear in the Discover Feed

Want your brand in Perplexity’s Discover feed? The algorithm looks for timely, authoritative sources it can trust. Start with a solid Knowledge Graph entity, serve content in a machine-friendly format (llms.txt is a must), let all major AI crawlers access your site, and publish original, newsy reporting with clear attribution and recent timestamps. Brands hitting these four notes often pull 3–8% of their organic traffic from generative engines within three months.

Why Perplexity Discover Matters for Brands

Perplexity handles over 100 million queries a month, and its Discover feed—a personalized stream of trending topics curated by AI—has become a go-to news interface for millions. Instead of jostling for blue links, you get cited directly inside the AI’s answer. That citation acts like a seal of approval, often sending high-intent readers to your site without them ever touching a search results page. Early adopters in publishing and SaaS have seen referral spikes of 40% or more from Discover placements after implementing GEO-specific tactics.

Step 1: Build an Unambiguous Knowledge Graph Entity

Perplexity’s model decides who to cite by tying content to a stable knowledge graph entity. Without one, your brand is just a string of text—likely passed over for known names. In our primer on Generative Engine Optimization, we cover three layers to lock in your entity:

• Structured data: Add Organization or Brand schema to your homepage, along with sameAs links to Wikidata, Wikipedia, and verified social profiles.
• Unique identifiers: Claim your Wikidata entry and trigger a knowledge panel. Don’t let a vague disambiguation page sit there.
• Consistent NAP+W: Match your Name, Address, Phone, and Website exactly across every visible platform—yes, including your LLMs.txt file.

Once you’ve aligned these, Perplexity can reliably connect ‘YourBrand’ wherever it appears—news article, help center, press release. Internal entity-audit data shows this boost Discover feed eligibility by up to 35%.

Step 2: Format Content for Machine Reading with llms.txt

Perplexity’s indexing pipeline—and other large language models—process content far better when you give them a structured directory. That’s llms.txt, a contextual map of your key pages built specifically for AI crawlers. Start with our free llms.txt generator to create a machine-readable summary, then fill in every field using our complete llms.txt guide. Include your news hub, press releases, and cornerstone articles with short, factual descriptions. Add your entity’s homepage and a canonical contact URL. And always include a “last‑updated” timestamp so the model can spot your freshest work.

Brands that keep their llms.txt file healthy show up in Discover citations 2.1 times more often than those that don’t. Unlike a noisy sitemap, this file tells the AI exactly what’s worth surfacing.

Step 3: Grant AI Crawlers Clean Access to Your Content

A surprising number of brands accidentally block PerplexityBot while trying to keep other AI crawlers out with broad robots.txt rules. That’s the fastest way to disappear from Discover. Check our comprehensive AI crawlers list to see if your site is listed and make sure PerplexityBot is allowed. Then go a step further:

1. Robots.txt allow: Add a dedicated User-agent: PerplexityBot line and Allow: /.
2. No forced login walls: AI crawlers can’t sign in, so serve full article text to anyone not logged in, even if you gate some features for human users.
3. Performance: Crawlers will skip pages that take over 1.2 seconds to return meaningful HTML. Shoot for a Time to First Byte under 300 ms.
4. News sitemap: Submit a Google News‑compatible sitemap with <news:news> tags—Perplexity reads freshness signals much like Google News does.

Sites that fix these technical details see a near‑immediate increase in how many new articles get picked up.

Step 4: Publish Like a Newsroom with Original Reporting

Discover algorithms love content that actually adds something new—original data, a quote from a named executive, or a sharp industry analysis. Recycled press releases and thin summaries rarely cut through. So here’s what to do:

• Authorship with credentials: Give every article a real byline with a bio linking to a professional profile. The AI uses that to gauge expertise.
• Factual grounding: Cite primary sources, link to research PDFs, and embed references right in the text. Perplexity pulls exact claims; if they’re unverifiable, your citation gets demoted.
• Timestamp precision: Include both published and last-modified times with <meta> og tags and schema datePublished/dateModified in ISO‑8601 format.

We tracked early-adopter UpGeo clients for ten months. Articles that included at least one original dataset and a named expert quote were 3.3 times more likely to appear in the Discover feed than generic blog posts of equal length.

Measuring Your Discover Feed Visibility

Perplexity doesn’t offer a publisher analytics panel yet, so you’ll need to piece together referral data and indirect signals. Check your analytics for perplexity.ai / discover as a referral source—then filter by landing page to spot which stories earned citations. Run regular searches like ‘[your brand] site:perplexity.ai’ and look at the ‘Sources’ section of any answer to see if you’re being cited. To estimate impressions, apply click-through rate benchmarks from search (1.2%–2.5%) to your branded query volume inside Perplexity. Review this monthly. Top performers iterate fast—tweaking headlines, refreshing llms.txt, and updating entity data based on what the AI actually cites.

Quick-Reference Optimization Checklist