Index: Who Controls AI Answers: The Complete Franchise Index · EPR Public Affairs Pillar Hub · EPR AI Communications Master Hub
Wire services and .gov own the facts. Think tanks own the framing.
An estimated top 5 sources supply ~65% of observed public-affairs answers, with .gov sources and wire services dominant.
Ballotpedia is the hidden winner. Structured, neutral, schema-tagged data on candidates, ballot measures, and election mechanics fills a gap government sources don't. No .gov source covers the same scope at the same depth. This is the Investopedia parallel for public affairs — definitional authority in a structured format is the most retrievable form of authority.
Cable news is the quiet loser. CNN, Fox, MSNBC are significantly under-cited versus print and wire on factual prompts. The engines treat video-first sources as lower-trust for retrieval. This is a structural disadvantage that no amount of reach compensates for.
Think tanks decide the framing layer. On "why did this happen" prompts where engines hedge factual claims, Brookings, AEI, and Pew supply most of the explanatory layer. The think tank that gets cited when the engine explains a policy outcome influences the framing of millions of AI-generated answers. That is not media relations. That is infrastructure.
Opinion, attribution of intent, and contested historical framing. Engines hedge on these prompts — and the think tanks they retrieve decide which framing wins. For public affairs practitioners, the implication is direct: the publications and institutions that earn think-tank and wire-service co-citation are the ones that shape AI answers on contested questions. PAC-owned domains rarely surface. The path to influence runs through Wikipedia accuracy and earned coverage in trusted publications.
Citation share modeled across four AI engines — ChatGPT, Claude, Perplexity, Google AI Overviews — and a fixed prompt set of 60+ queries. Sources tagged on the five-tier Retrieval Hierarchy: T1 Government & Academic · T2 Encyclopedic · T3 Publisher & Trade Press · T4 Community Platforms · T5 Brand-Owned. Estimates directional and date-stamped.
Which sources do AI engines cite most for politics? Reuters, AP, federal .gov sites, Wikipedia, NYT, WaPo, BBC, Pew, Ballotpedia, and Brookings. Wire services and .gov sources dominate factual prompts.
How do think tanks influence AI political answers? On "why did this happen" prompts where engines hedge factual claims, Brookings, AEI, and Pew supply most framing-layer citations.
Can political brands influence their AI citation share? Indirectly. The path runs through Wikipedia accuracy and earned coverage in trusted publications.
The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.
USAFacts launched The Data We Depend On June 9, 2026 — mobile OOH in NYC, LA, DC, Miami plus an open letter to Congress on federal data infrastructure.
Healthcare decisions have moved into AI faster — and at higher stakes — than almost any other category. The prompts patients actually run are not abstract. "Best cancer hospital in the US." "Should I trust Mayo?" "Best hospital for pancreatic cancer?" "Top orthopedic surgeon in N
FinTech startups that fail to define their category early get defined by the market — usually wrong. Kevin Mercuri, CEO of Propheta Communications, on why early-stage PR is foundational infrastructure, not a late-stage growth tactic.
EPR publishes the data every week.
Free. Weekly. Unsubscribe anytime.