Why do expert quotes matter to AI engines?

Because grounded answers attribute claims to people. When Gemini 3 or Perplexity composes an answer about, say, supply-chain software trends, the passages that survive grounding are specific statements attached to identifiable humans with credentials. “Dr. A, professor of logistics, notes that...” is exactly the shape engines prefer. Your executives and specialists need to exist in the corpus as quoted experts, not just as bylines on your own blog. This complements the entity work in author entities and the PR motion in digital PR for GEO.

What replaced HARO?

After HARO shut down, the connective tissue between journalists and sources fragmented into several channels:

  • Qwoted and Featured: structured query-response platforms where journalists post questions and experts pitch answers.
  • Direct relationships: curated lists of sector journalists built from bylines, nurtured over months — slower but far higher conversion.
  • Community channels: Slack and Discord groups where reporters request sources, plus social call-outs.
  • Reactive monitoring: watching the news cycle for stories where your data or expertise is the missing quote.

The platforms commoditized fast, so response quality is the differentiator: most pitches are generic, and journalists — like grounding systems — discard them.

What makes a quote citable by AI, not just printable?

  1. One specific claim per quote. Numbers, mechanisms, predictions with timeframes. Vague agreement (“AI is changing everything”) is never extracted.
  2. A consistent title and bio. The expert's role should be phrased identically everywhere so the person entity consolidates; inconsistency weakens attribution confidence.
  3. A standalone sentence. If the quote only makes sense inside the article's framing, engines cannot lift it into a different answer.
  4. Alignment with your owned content. The quote should match what your site says, because grounding cross-checks the person against their domain.

How do you operationalize outreach?

Build a bench: three to five named experts, each with a one-paragraph canonical bio, headshot, and a list of five topics on which they can comment with genuine authority. Assign response SLAs — journalist queries expire in hours, and Perplexity's real-time retrieval means a story published today is in answers tonight. Track every placement: publication, URL, quote text, date. This register becomes an asset for brand mention monitoring and for measuring whether quoted claims actually surface in answers.

What does the 2026 data context say?

The Ahrefs February 2026 study across 300,000 keywords found that AI Overviews cut position-1 CTR by about 58% and leave roughly 68% of sessions zero-click. Earned media is one of the few channels that reaches buyers inside those zero-click answers: the article quoting your CTO is the passage Perplexity cites at sentence level. With Perplexity past 1 billion monthly queries, each strong placement is retrieved far more often than its pageview count suggests.

What pitfalls should you avoid?

First, mass-produced AI-written pitches: journalists can detect them, platforms are banning them, and even when published they contain no extractable claim. Second, title inflation that does not match the rest of the web — if the quote says “Chief Scientist” but LinkedIn says “Product Manager”, attribution confidence collapses. Third, chasing volume over relevance: a quote in a general-interest outlet on an off-topic question adds little to your semantic territory. Fourth, treating outreach as one-off: the compounding asset is the relationship and the expert's growing footprint, which takes quarters. Finally, never fabricate credentials or commentary on topics outside your expertise — one debunked quote contaminates the expert entity everywhere it appears, because grounding treats credibility as a property of the person, not the article.

Fold placements into your monthly measurement: add questions your experts comment on to the fixed 50-100 question set and check, with screenshots, whether answers attribute claims to them by name. That attribution rate is the true KPI of expert-quote work.

Frequently asked questions

Is HARO completely gone?
The original HARO service shut down and the space fragmented into platforms such as Qwoted and Featured, plus direct journalist relationships and community channels. The underlying practice — experts answering journalist queries — is unchanged and still effective.
What kind of quote gets cited by AI engines?
A specific, standalone claim with a number, mechanism or dated prediction, attributed to a named person with a consistent title. Vague commentary is rarely extracted or cited.
How many experts should a company maintain for outreach?
Three to five is practical for most firms: enough topic coverage to match journalist demand, few enough to keep bios, titles and messaging perfectly consistent across every appearance.