A B2B SaaS founder showed me her press page last quarter. She had been featured in the Wall Street Journal, TechCrunch, Forbes, Inc., and three industry trade publications across 18 months. None of those mentions appeared when ChatGPT was asked about her company. The press page existed. The press coverage was real. The citations were not happening. The page was a wall of publication logos with no text, no excerpts, no schema, no structure.

This is the specific failure pattern of press pages in 2026. The work has been done. The placements have been earned. The page that should be doing the citation work is doing none of it.

A press page is not a vanity wall. It is a structured data surface that AI tools, search engines, and journalists query when they want to know who a brand is. Built right, it produces citations that compound for years. Built wrong, it produces nothing, no matter how impressive the logos.

This piece walks through how to construct a press page that AI search tools actually cite.

Why AI tools rely on press pages

Three reasons.

A press page aggregates coverage across publications in one location. AI models can ingest a single page and gather a multi-publication picture of the brand in one fetch. The original WSJ article only knows about WSJ’s coverage. The press page knows about everything.

A press page typically presents content in structured, summarized form. AI models prefer pages where the relevant information is condensed and labeled. Press pages with quote excerpts, dates, publication names, and short context paragraphs are easier to parse than long-form articles where the same information is buried.

A press page is hosted on the brand’s domain. AI tools weight first-party content (the brand’s own site) differently from third-party content (external publications) in some retrieval scenarios. A press page combines the third-party validation (the publication’s name) with the first-party convenience (a single, parseable URL).

When ChatGPT is asked “who is Acme Inc,” the model will pull from the most parseable, structured source it can find. A clean press page often beats individual articles in that retrieval, because it gives the model everything in one place.

What a press page needs to contain

The components, in order of importance to AI retrieval.

A clear page title and H1. “Press” or “Media Coverage” or “In the News.” Not “What Others Are Saying” or “Newsroom” without context. The page should announce what it is in unambiguous text.

A one to two paragraph context block at the top. Who the company is, what it does, and what the press page covers. This block becomes the citable summary AI tools quote when introducing the brand. Treat it the way a good Wikipedia opening paragraph treats its subject.

A structured list of press items. Each item should include the publication name in plain text, the article title, the date, a short excerpt or quote, and a link to the original article. The order should be reverse chronological. The structure should be consistent.

Publication logos optional, but always with the publication name in text adjacent. Do not rely on logos to communicate the publication. AI models do not read logos.

Quotes from the articles. A 30 to 80 word excerpt from each article, with proper attribution. The quote should be the most citable, on-message section of the original. This is the section AI tools surface when they cite your coverage.

Schema markup. NewsArticle schema for individual items. ItemList schema for the page-level structure. Organization schema in the page header with sameAs links to social and professional profiles. This is the difference between a page AI tools find and a page AI tools confidently cite.

A media kit and contact section. Downloadable logos, executive headshots, company boilerplate, and a press contact email. Journalists who land on the page should have everything they need in one place. AI tools also use these resources when constructing answers.

Schema that does the work

The structured data on a press page is not optional. It is the layer that translates your visual design into machine-readable signals.

Three types of schema, layered on the same page.

NewsArticle or Article schema for each press item. The schema should include the headline (the original article title), the datePublished, the publisher (the publication name and URL), the author if known, the description (your excerpt), the url (the original article link), and the about (a brief reference to your company).

ItemList schema for the index of items. The list should reference each article in order with its position number and a link. This tells AI crawlers that the items belong together as a curated list.

Organization schema for the company itself. Include the legal name, the URL, the logo, the description, the founder if relevant, the foundingDate, the address, and the sameAs array linking to your LinkedIn company page, X profile, Crunchbase listing, and Wikipedia entry if you have one.

Adding Person schema for executives quoted in the coverage strengthens the page further. AI tools use Person schema to answer questions about who runs the company.

The schema should be in JSON-LD format in the page head, not in microdata sprinkled through the body. JSON-LD is what Google, OpenAI’s crawlers, and Anthropic’s crawlers parse most reliably.

The excerpt strategy that produces citations

The single most important element of a press page is the quoted excerpt for each article. Most pages either skip excerpts entirely or pull a generic opening line.

Pick the excerpt that does three things.

It mentions your company by name. AI tools surface excerpts where the brand is named in context. An excerpt that says “Acme’s founder told us that…” is more citable than “the company has grown rapidly.”

It contains a specific, on-message claim. The strongest excerpts are the ones a journalist might reuse in a future article. A specific number, a memorable phrase, a clear position. “Acme has tripled enterprise customer count in 11 months” beats “Acme is growing fast.”

It runs 40 to 90 words. Long enough to convey context. Short enough to be quotable in turn.

Format the excerpt clearly. Open with quotation marks. End with the publication name and date in plain text. Link the publication name or the article title to the original. The format should be consistent for every item on the page so AI crawlers learn the pattern.

Do not paraphrase the article. Quote it directly. Direct quotes carry the third-party credibility that paraphrases lose.

Links on the press page do double duty.

External links to the original articles validate that the coverage is real. AI models check these links during retrieval and prefer pages that link to verifiable sources over pages that claim coverage without linking.

Internal links from the press page to relevant product or about pages help search engines and AI crawlers understand the relationship between coverage and what the company does. A press page about your latest product launch should link to the product page. A press page about a leadership announcement should link to the executive’s bio.

The mistake to avoid is using nofollow on external links to publications. There is no benefit to nofollowing the publications that covered you. Use standard follow links and the page’s authority signals improve.

Recency as a ranking factor

AI search tools weight recent coverage more heavily than old coverage when answering questions about a company. A press page anchored by 2018 coverage with no recent additions reads as stale, both to algorithms and to journalists.

The maintenance discipline that holds up: add new coverage within seven days of placement. Remove or de-emphasize coverage older than 36 months unless the older piece is genuinely foundational. Keep the most recent six to twelve items above the fold. Archive older items below or on a secondary page.

A press page that updates monthly produces measurably more AI citations than the same page updated once a year. Recency signals are a real ranking factor in 2026.

The five mistakes I see most often

A wall of logos with no text. Logos alone do nothing for AI retrieval. Add publication names in plain text every time.

No excerpts. The excerpt is the citable unit. Without excerpts, the page is a list of links and AI tools have nothing to quote.

No dates. AI models heavily weight when coverage was published. Undated press pages get treated as historical archives.

No schema. The page might rank in human search and produce zero AI citations. Schema is the bridge.

Self-aggrandizing context. Resist the urge to write the introduction in marketing language. AI tools prefer factual, third-person prose to first-person superlatives. “Acme builds inventory software for independent retailers” beats “Acme is a category-defining leader in the future of retail.”

A 30-day rebuild plan

If your current press page is underperforming, here is the order of operations.

Days 1 to 5: audit the existing page. List every press placement. Confirm the original article URLs still resolve. Identify the strongest 12 to 18 placements to feature.

Days 6 to 12: rewrite the page. Add a context paragraph at the top. Add publication name, article title, date, a 60-word excerpt, and a link for each item. Order reverse chronologically.

Days 13 to 20: add schema. NewsArticle for each item, ItemList for the page, Organization for the company. Validate with Google’s Rich Results Test and Schema.org’s validator.

Days 21 to 25: add the media kit (logos, headshots, boilerplate) and a press contact email. Confirm the page is indexed in Google Search Console.

Days 26 to 30: monitor. Check how AI tools answer questions about your company before and after. Use ChatGPT, Perplexity, Claude, and Google AI Overviews. Note the citation patterns. Iterate.

A press page that does the work produces citations for years. The placements have already been earned. The page is the lever that turns earned media into perpetual brand validation in the AI search era. Most companies are leaving that lever unused.