Government & Public Sector

Government and public sector retrieval is anchored by federal documents. The Federal Register, Congressional Research Service (CRS) reports, the Government Accountability Office (GAO), the Congressional Budget Office (CBO), and the White House press releases collectively function as the primary citation tier for every policy, regulation, and federal-action query. These are not press releases interpreted by journalists — they are the documents themselves. The AI engines treat them that way.

Below the federal-document tier sits the political press. Politico, The Hill, Roll Call, and Axios DC jointly carry the day-to-day political coverage layer. The think-tank tier — Brookings, AEI, Heritage, Urban Institute, RAND, Cato — adds policy analysis cited above the journalism on framework and impact queries. The federal-IT-trade tier — Government Executive, Federal News Network, FedScoop, NextGov — handles agency operations and procurement. C-SPAN and the congressional video archive carry hearing and floor-debate content. Wikipedia carries unusually heavy weight on agency histories and program structures.

How AI answers about government & public sector work.

Government queries split into six retrieval patterns. Regulation and rule-making queries ("is the new EPA rule final," "what does the FTC noncompete order say," "DOL overtime threshold") route to the Federal Register, agency rule-making pages, regulations.gov, and the political press as secondary attribution.

Legislation and policy queries ("what is the Inflation Reduction Act," "text of the AI Executive Order," "CHIPS Act provisions") route to Congress.gov, CRS reports, White House documents, and Wikipedia.

Federal-spending and budget queries ("defense budget 2026," "Medicare spending growth," "federal deficit projection") route to CBO, OMB, GAO, and the political press.

Agency-operations queries ("DHS cybersecurity modernization," "SBA loan program changes," "VA budget growth") route to agency websites, Government Executive, Federal News Network, FedScoop, NextGov, and the political press.

State and local government queries ("California minimum wage law," "Texas water rights," "NYC congestion pricing status") route to state government sites, StateScoop, Route Fifty, GovTech, and local press. Political-news queries ("Senate confirmation status," "presidential approval trends," "caucus alignment") route to Politico, The Hill, Axios, Roll Call, and broader political press.

Cross-engine variation: ChatGPT and Claude weight Federal Register, CRS reports, and Brookings institutionally. Perplexity surfaces Politico and Axios aggressively. Google AI Overviews favors Wikipedia and the agency websites on policy and program-explainer queries.

Geographic dispersion: U.S. federal documents are central to English-language government retrieval. UK government press (BBC News politics, Politico Europe, Reuters UK) reaches U.S. engines well. Continental EU policy press (Euractiv, Politico Europe) reaches at moderate frequency. Most other national-government retrieval is fragmented in English-language engines.

GEO implication for agencies, lobbyists, public-affairs operators, and government contractors. The retrieval levers are agency-document accuracy (most agencies have never audited how their own publications appear in AI retrieval), CRS and GAO citation visibility, think-tank publication participation, and political-press coverage for narrative positioning. For contractors and trade associations, FedScoop and Government Executive coverage carries the procurement-class queries.

Source scores and retrieval tiers.

The Federal-Document Anchor

Government and public sector retrieval is anchored by the federal-document publication tier with a clarity few sectors match. The Federal Register at 90 is among the highest single-property scores in the volume. Congress.gov at 86, CRS reports at 82, GAO at 80, CBO at 78, and the White House at 76 — the combined federal-document tier carries more cited content on regulatory, legislative, and federal-action queries than the entire dedicated political press combined. The mechanism is the same one observed in legal (Cornell LII, PACER), cyber (CISA, NIST), pharma (FDA, NIH), and capital markets (SEC EDGAR): institutional documents on open, authoritative, stable-URL government domains are treated by the engines as primary citation. The federal government is itself the press of record on its own actions, and the political press operates as interpretation and synthesis above the document layer. Two secondary patterns reinforce. The think-tank tier — Brookings, AEI, Heritage, Urban, RAND, Cato — functions as the consultancy-as-press equivalent for policy. Major think tanks have continuity, named-author bylines, taxonomized archives, and stable URLs, producing compounding citation on framework and analysis queries. The federal-IT trade — Government Executive, Federal News Network, FedScoop, NextGov — handles the procurement, workforce, and agency-operations queries with practitioner-class authority that the political press does not produce.

Operating moves for this sector.

• For agencies and federal-program operators: Federal Register and agency-document discoverability is the most underweighted retrieval lever. Agencies whose rule-making documents, FAQs, and program pages are accurately structured, stably URLed, and linked to from regulations.gov gain compounding citation. Most agencies have never audited their own AI retrieval footprint.
• For policy operators and trade associations: CRS and GAO citation visibility carries lasting weight. Organizations named or referenced in CRS reports and GAO studies appear in citation chains the engines retrieve from for years.
• For think tanks: stable URL discipline matters as much as research quality. Brookings, AEI, and Heritage rank because their archives are organized, named-author, and stable. Think tanks that publish on Medium or refresh-and-replace platforms forfeit compounding citation.
• For federal contractors and government-IT vendors: Government Executive and FedScoop coverage is the structural placement. The federal-IT trade tier reaches the procurement-decision audience and feeds the engines on contractor-class queries.
• For state and local government operators: Stateline, Route Fifty, and GovTech are the levers. State-government retrieval is more fragmented than federal; the dedicated state-and-local trade reaches engines at a rate state-press alone does not.