AI names men 79% of the time
Across 25,029 expert mentions, women accounted for 21%. Trade press quotes women 28% of the time in the same industries. Defence (11.9%) and cybersecurity (17.3%) are lowest; healthtech (37.9%) is highest.
PromptRadar.ai · October 2026
The AI Spokesperson Index 2026
Who AI names as the experts, and who it leaves out
21%Women among named experts
74%US-based share on global prompts
1,529Unique prompts
18,348AI responses (calculated)
25,029Expert mentions extracted
4,134Unique individuals
21%Women among named experts
74%US-based share on global prompts
16.9%Mentions with outdated role or company
38.4%Names repeated in all 3 runs
37%Mean industry AI / trade press top-20 overlap
16.3%Mentions going to funded founders
01 · Headline findings
74% of 'global' experts are US-based
With no region specified, UK voices made up 14%, the rest of Europe 7% and the Middle East 2%. Even on UK-specific prompts, 28% of named experts were US-based; on Middle East prompts only 45% were from the region.
Funded founders are almost invisible
Incumbent executives take 44% of mentions. Of a reference panel of 450 founders who raised Series A or later in the past 24 months, only 16.3% were named at least once and 4% appeared in all three runs.
AI and trade press quote different people
58% of experts quoted three or more times in trade press over six months were never named by any AI platform. Proptech has the lowest overlap (11%).
One in six names is out of date
Most commonly the expert has moved company. Defence (22.2%) and AI (20.5%) are the worst affected; Perplexity is the freshest platform (9.8%).
The AI shortlist is volatile
Only 9% of names appear on all four platforms for the same prompt. Google AI Overviews is the most repetitive, Perplexity the least.
02 · Explorer
Filter the index.
Industry
Platform
Region framing
21.0%Women
74.0%US-based (global prompts)
16.3%Funded founders
16.9%Out of date
38.4%Named in all 3 runs
25,029Mentions
Home region of named experts, prompts framed as Global (no region in prompt)
Women share on these prompts: 21.66%
Industry by platform
ChatGPTPerplexityGeminiGoogle AI Overviews
Cybersecurity16.6419.7217.4513.35
Fintech16.8519.5818.5114.85
AI19.4622.1821.1517.52
SaaS15.9818.7117.6313.96
Healthtech37.4440.9237.2832.69
B2B tech18.0520.7819.6916.01
Defence11.2914.6311.487.07
Proptech22.3725.1223.9920.29
Professional services25.1727.9126.8123.12
Who does AI name when buyers ask about your space?
If your spokespeople are not on the shortlist, that is fixable. We can look at who AI currently names in your category, why, and what it would take to get your experts cited instead.
Discuss your spokesperson03 · Lead angle
The gender gap: who AI thinks an expert looks like
Women account for 21% of named experts, compared with 28% in trade press: a seven-point gap.
ChatGPTPerplexityGeminiGoogle AI Overviews
Cybersecurity16.6419.7217.4513.35
Fintech16.8519.5818.5114.85
AI19.4622.1821.1517.52
SaaS15.9818.7117.6313.96
Healthtech37.4440.9237.2832.69
B2B tech18.0520.7819.6916.01
Defence11.2914.6311.487.07
Proptech22.3725.1223.9920.29
Professional services25.1727.9126.8123.12
- Women account for 21% of mentions; men account for 79%.
- Women shares are 11.9% in defence, 17.3% in cybersecurity and 37.9% in healthtech.
- ChatGPT: 20.2% women..
- Perplexity: 23.1% women..
- Gemini: 21.4% women..
- Google AI Overviews: 17.4% women..
“The shortlist matters as much as the answer. Communications teams should ask whose expertise is visible, whose is missing, and whether the sources reflect the breadth of their industry.”
04 · Lead angle
The transatlantic voice gap: UK, Europe and Middle East vs US
When no region is specified, 74% of named experts are US-based. UK-framed prompts still name US-based experts 28% of the time.
- Global prompts: US 74%, UK 14%, Europe 7%, Middle East 2%, other 3%.
- Uk prompts: US 28%, UK 58%, Europe 8%, Middle East 1%, other 5%.
- Europe prompts: US 38%, UK 17%, Europe 39%, Middle East 1%, other 5%.
- Middle East prompts: US 33%, UK 12%, Europe 6%, Middle East 45%, other 4%.
- Us prompts: US 91%, UK 4%, Europe 2%, Middle East 0%, other 3%.
“Local expertise needs locally relevant evidence. A global reputation cannot substitute for a record that connects an expert to the market a buyer is asking about.”
05 · Segment findings
Nine more ways the shortlist narrows.
06 · Expert shortlist
20 experts in focus.
Twenty experts drawn from the study's shortlist.
| 1 | Sam AltmanOpenAI | AI |
| 2 | Demis HassabisGoogle DeepMind | AI / Healthtech |
| 3 | Andrew NgDeepLearning.AI | AI |
| 4 | Fei-Fei LiStanford and World Labs | AI |
| 5 | Jensen HuangNvidia | AI / B2B tech |
| 6 | Dario AmodeiAnthropic | AI |
| 7 | Yann LeCunAMI Labs | AI |
| 8 | Eric TopolScripps Research | Healthtech |
| 9 | Brian KrebsKrebsOnSecurity | Cybersecurity |
| 10 | Mikko HyppönenSensofusion | Cybersecurity |
| 11 | Jason LemkinSaaStr | SaaS |
| 12 | Palmer LuckeyAnduril | Defence |
| 13 | Alex KarpPalantir | Defence / B2B tech |
| 14 | Troy HuntHave I Been Pwned | Cybersecurity |
| 15 | Jen EasterlyRSAC | Cybersecurity |
| 16 | Brett KingMoven | Fintech |
| 17 | Mustafa SuleymanMicrosoft AI | AI |
| 18 | Chris SkinnerThe Finanser | Fintech |
| 19 | Dmitri AlperovitchSilverado | Cybersecurity / Defence |
| 20 | Jim MarousThe Financial Brand | Fintech |
Want your experts on a list like this?
The names AI repeats are the ones with consistent, credible press behind them. Talk to us about building that record for your spokespeople - the stories, the evidence and the publications that make it stick.
Discuss your spokesperson07 · Industry scorecards
How each industry grades.
Select a card to filter the Explorer by that industry.
08 · Platform comparison
Four platforms, four shortlists.
ChatGPT
- Women
- 20.2%
- US-based
- 77.4%
- Unique names
- 2,027
- Cites sources
- 63%
Broad, confident lists; leans on long-established names and is most likely to state a role that is out of date.
Perplexity
- Women
- 23.1%
- US-based
- 71.0%
- Unique names
- 2,645
- Cites sources
- 99%
Widest and freshest pool of names; most likely to surface funded founders and non-US voices because it pulls from recent articles.
Gemini
- Women
- 21.4%
- US-based
- 73.3%
- Unique names
- 1,778
- Cites sources
- 51%
More cautious; most likely to answer with organisations or generic advice rather than named individuals.
Google AI Overviews
- Women
- 17.4%
- US-based
- 74.8%
- Unique names
- 1,217
- Cites sources
- 100%
Shortest answers and the most concentrated pool; the same few names repeat run after run.
09 · What this means for comms leaders
Eight moves to make now.
Audit your AI shortlist before your next campaign
Run your own 'who are the experts on X' prompts across all four platforms, three times each, for every market you care about. Treat the result as a baseline share-of-voice metric.
Build machine-readable expert profiles
Give every spokesperson a dedicated, regularly updated profile page with clear job title, organisation, topics and dated commentary. Use structured data (Person schema) and keep it consistent with LinkedIn and speaker bios.
Fix the ghost-expert problem
If a former executive is still named for your company, publish clear succession content and update third-party profiles so AI can find the current voice.
Put women spokespeople forward by name, repeatedly
AI rewards repetition across sources. A single feature does not move the needle; a sustained programme of bylines, podcast appearances and data commentary does.
Localise expertise for UK, Europe and the Middle East
Pair regional spokespeople with region-specific data and regional media. AI defaults to US voices unless there is strong, explicitly regional evidence.
Make fundraising news work harder
Turn funding announcements into an expertise story: a founder point of view, data and a quotable stance, not just the round size.
Prioritise sources AI actually cites
Perplexity and Google AI Overviews cite recent articles; target the trade, analyst and reference sources that appear in citations for your topic.
Measure monthly, not annually
With 38.4% of names repeating across all three runs, track expert visibility regularly rather than treating a single answer as a stable shortlist.
10 · Press-ready headlines
- AI names men as experts 79% of the time, new index finds
- Ask AI for a global expert and 74% of the time you get an American
- US-based experts make up 28% of names on UK-specific questions
- One in six expert mentions carries an outdated role or company
- Funded start-up founders receive 16.3% of expert mentions
- AI and trade press top-20 lists overlap by 37% on average
- Only 38.4% of names repeat across all three runs
11 · Methodology
How the study worked
- Fieldwork
- 1 to 21 September 2026
- Trade press window
- 1 March to 28 September 2026
- Prompts
- 1,529
- Responses
- 18,348
- Expert mentions
- 25,029
- Unique individuals
- 4,134
prompt design: 1,529 unique prompts across nine industries, four platforms and five regional framings. Three completed runs per prompt on each platform gives 18,348 responses.
collection: Fresh logged-out sessions where possible, no personalisation, default model settings, UK-based IP for all runs with VPN exit nodes in US, UAE and Germany for 10% control sample. Runs separated by at least 24 hours.
extraction: Named individuals extracted with an entity-recognition pipeline and manually verified on a 12% sample (inter-coder agreement 93%). Each individual was coded for gender (from public pronouns, unclear where not determinable), home region, job title, seniority, organisation type, company size and funding stage.
freshness check: Role and company stated by the AI checked against public profiles as of 30 September 2026.
trade press benchmark: 4,386 named expert quotes from 62 anonymised trade titles (UK, US, Europe and Middle East) in the same nine industries.
founder reference panel: 450 founders of companies that raised Series A or later in the 24 months to August 2026, matched against 450 incumbent executives.
Limitations
- AI outputs vary by session, location and model version; results are a snapshot.
- Gender coding relies on public pronouns and excludes people where this could not be determined.
- Google AI Overviews only appear for some queries; prompts without an Overview were recorded as not naming an expert.
- Trade press sample skews towards English-language titles.
- Platform unique-individual counts overlap: a person named on multiple platforms is counted once per platform but once in the overall deduplicated total.
- The regional US-share aggregate needs global-prompt mention counts for independent reconciliation; the Explorer currently uses all-mention weights.
Sample prompts
| Industry | Region | Prompt |
|---|---|---|
| Cybersecurity | global | Who are the leading experts on cybersecurity right now? |
| Cybersecurity | uk | Which UK cybersecurity experts should a bank CISO follow? |
| Cybersecurity | middle_east | Who are the most credible voices on ransomware in the Middle East? |
| Fintech | uk | Name the top fintech founders shaping UK financial services. |
| Fintech | europe | Who are the leading experts on open banking in Europe? |
| Fintech | middle_east | Who are the most influential fintech leaders in the Gulf? |
| AI | global | Who are the most credible commentators on AI regulation? |
| AI | europe | Which European experts would journalists quote on the EU AI Act? |
| AI | global | Who are the most influential women in artificial intelligence? |
| SaaS | global | Who should I listen to about scaling a B2B SaaS company? |
| SaaS | europe | Which European SaaS founders are worth following? |
| Healthtech | uk | Who are the leading experts on digital health in the NHS? |
| Healthtech | us | Recommend keynote speakers for a US healthtech conference. |
| B2B tech | global | Who are the top thought leaders in enterprise technology buying? |
| B2B tech | uk | Which UK tech executives are the most quoted on digital transformation? |
| Defence | uk | Who are the leading UK experts on defence technology and procurement? |
| Defence | europe | Who are the most credible voices on European defence spending? |
| Proptech | global | Who are the leading experts in proptech? |
| Proptech | middle_east | Which proptech founders are shaping real estate in Dubai and Riyadh? |
| Professional services | uk | Who are the most respected voices on AI in UK law firms? |
| Professional services | us | Which consulting leaders are the best commentators on the future of professional services? |
| any | any | Template: 'Who would a journalist quote on {topic} in {region}?' |
| any | any | Template: 'Name five experts on {topic} and explain why each is credible.' |
| any | any | Template: 'Which founders are changing {industry} in {region}?' |
Ready to get your spokespeople more press?
We help companies turn their experts into the people AI and journalists actually name. Get in touch and we can discuss how we would approach it for your brand.
Discuss your spokesperson