How to Get Executives Quoted in the Outlets AI Assistants Cite

An executive quote most often reaches a ChatGPT, Claude or Gemini answer inside an article the assistant already retrieves for your category's questions. The assistant cites the article's URL, and the executive's name and argument come along only if the reporter wrote them in. So the job is three tasks done in order: find the outlets those answers cite today, give their reporters something they can attribute, and check whether each placement turns into a citation.

What the citation studies measured, assistant by assistant

For B2B AI search visibility questions about citations, sources, earned media and PR, the most detailed public data we found is Muck Rack's What Is AI Reading?, first run in July 2025 and updated three times since. It uses realistic consumer prompts across industries, so B2B teams should treat its technology and industry-trend findings as the closest proxy for their buyers.

Earned runs far above journalism because Muck Rack also counts third-party corporate content, academic research, government and NGO sites, encyclopedic sites and social platforms as earned.

The October update shows the assistants drifting apart. Journalism fell from 28.1% of ChatGPT's citations in January to 22% by September, while Claude's share moved from roughly 28% to 24% and Gemini's from 27.6% to 23.6%. We found no journalism share for Perplexity in the editions we read. The October edition did test its consistency: when researchers ran the same prompt twice in one day, Perplexity repeated 94% of its cited domains, against 41% for ChatGPT and 45% for Gemini.

Depth differs too. In May 2026, ChatGPT cited sources in 96% of responses at about five each, Gemini in 82% at about eight, and Claude in 55%, averaging 13 when it cited at all. Forbes was the only journalism outlet in more than one provider's top five.

Four findings bear directly on executive visibility:

• Question type: industry-trend prompts cited journalism in 46% of responses in May 2026, more than twice the rate for how-to and comparison prompts, and drew on smaller outlets and newer articles. Commentary on where a market is heading fits that question type.
• Concentration: in the December 2025 report, most AI citations for a given brand came from about 20 outlets, and each brand's mix was different.
• Mismatch: the same report compared the journalists most pitched on Muck Rack with those AI cited most for a particular brand and found an average overlap of 2%.
• Trade press: technology prompts cited journalism 24.2% of the time in May 2026, with trade publications ahead of general-interest news.

How executive commentary becomes a citable source

Brands get cited in AI-generated answers through executive thought leadership when the commentary appears in earned media from authoritative sources the assistant already reads, and the generative engine optimization research behind 2025 and 2026 practice suggests why quotes and numbers help.

A quote that pairs a named speaker with a figure gives the model a specific claim to lift with attribution. The original GEO paper by Aggarwal and colleagues, posted on arXiv and accepted to KDD 2024, tested content edits against a benchmark of queries. Adding quotations from credible sources, adding statistics and citing sources were its top methods, raising a source's visibility in the generated answer by 30–40% in relative terms on its position-adjusted word count measure. On Perplexity.ai, adding quotations led at 22%. The main experiments ran a GPT-3.5 answer engine over the top five Google results, so the paper shows which text a model uses after retrieval; brand citations in current assistants need their own measurement.

Press releases show the same lean toward numbers: in its December 2025 findings, Muck Rack reported that cited releases carried about twice as many statistics, on average, as uncited ones.

Reporters ask for the same raw material. In the State of Journalism 2026 survey from Muck Rack (897 responses in the final analysis, collected January 30 to March 2, 2026), 70% of journalists wanted clear relevance to their beat, 58% wanted interview access to relevant sources and 40% wanted original data or research. Pre-written quotes drew 14%, and 88% said they immediately disregard pitches that miss their coverage area. The executive earns the quote by taking the interview with a number in hand.

Which thought leadership formats feed LLM sourcing

On the evidence so far, executive thought leadership works as an LLM sourcing tactic for GEO roughly in proportion to how independent the page carrying it is.

• Reported articles and interviews sit inside the earned majority that drives discovery questions.
• Paid contributor columns trail editorial. In the October 2026 analysis of ChatGPT's Forbes citations, editorial content outperformed paid Forbes Councils content by about 1.7 to 1.
• Owned posts matter for fact-finding questions. The December 2025 report found earned media drove discovery prompts while the brand's own channels were usually the cited source for specific facts.
• Press releases made up 1.1% of citations in May 2026, down from about 6% in December 2025, and appeared most in industry-trend answers.
• For executive LinkedIn posts, we found no separate figure in the editions we read.

So publish the data behind an executive's claim on a page you own, where fact-finding prompts and reporters can find it, and keep the executive's name, title and company identical across bio, press kit and profiles; our entity consistency audit explains why contradicting facts cost citations.

A method for getting executives quoted where assistants read

To get executives quoted as sources that appear in AI-generated answers, and turn those quotes into brand visibility, run three steps in order: map, pitch, test.

List the outlets already cited for your category prompts

1. Write 25 to 40 prompts your buyers would ask, weighted toward industry-trend questions.
2. Run each in ChatGPT, Claude, Gemini and Perplexity at least twice. With repeat rates as low as 41%, one run misses sources.
3. Log every cited URL and tag its type with Muck Rack's categories (journalistic, owned, press release and so on).
4. Roll the URLs up to outlet and byline and count them. Expect a short head, in line with the 20-outlet finding.
5. Check each outlet's robots.txt before pitching. In the October 2026 analysis, 81% of major publishers blocking Claude's search crawler received almost no Claude citations, while just 12% of publishers allowing it saw similarly low citation levels.

A hypothetical payroll software vendor runs 30 prompts through four assistants twice each, 240 answers in all. Every figure below is invented to show the layout.

Pitch with data a journalist can attribute

The pitch carries four things:

• One finding from your own data, with its population, sample size and date in the same sentence.
• A two-line method note and an offer to share the dataset.
• The executive's availability for an interview, with one line on why they can explain the finding.
• The tie to that outlet's readers. In the same survey, 78% of journalists said a pitch feels genuinely relevant when it directly affects the community their audience belongs to.

Send it to the bylines from your map. Given the 2% overlap, expect many of them to be missing from your current media list.

Test whether each placement becomes a citation

Log each placement's URL, publish date, outlet and target prompts, then rerun those prompts at one week, one month and three months. About 4% of ChatGPT and Claude citations in the December 2025 report pointed to articles from the previous seven days, the highest rate of any period, and in May 2026 journalism citations peaked in the first month after publication and dropped through month six.

Score three outcomes separately, because they fail separately: the article URL is cited, the brand is named in the answer, and the executive is named or quoted. Fold all three into your AI share of voice baseline so one good week does not pass for a trend. An outlet that returns nothing after two cycles moves down the list, and one that converts gets the next exclusive. The on-page side is covered in how to get cited by ChatGPT and what GEO is. For a first read on where your site stands, try the AI visibility checker, or see our AI search optimization service for outside help.

Sources

• Muck Rack: What Is AI Reading? December 2025
• Muck Rack: What Is AI Reading? May 2026
• Muck Rack: October 2026 edition summary, updated October 28, 2026
• Muck Rack: December 2025 findings, updated February 16, 2026
• Muck Rack: State of Journalism 2026
• arXiv: GEO: Generative Engine Optimization, Aggarwal et al., KDD 2024