Report · Coverage

The Verifiability Frontier

A map of what ASURIQ’s 21 evidence databases cover well, and where verification demand plausibly exists with no comparable public database to serve it yet.

Published 30 July 2026·First report
1. Why this report exists

A coverage gap is a finding, not a defect.

When a claim can’t be checked against any adequate public database, that absence is itself information about the state of public information — not a failure of the engine, and not evidence the claim is false. Publishing where public data cannot answer a question is a legitimate finding in its own right, and it is information nobody else is positioned to publish: this report only exists because ASURIQ sits where verification demand meets database absence, across a wide enough span of everyday claims to see the pattern.

2. Methodology note

Qualitative reasoning, not a market-sizing exercise.

This report reasons directly from the coverage categorization published on the methodology page — the same 18 free-tier databases plus 3 commercial sources every ASURIQ verdict draws on. ASURIQ does not have access to live query-volume analytics, so every frontier gap below is a reasoned, qualitative observation: for each domain category the 21 sources cover well, which adjacent domains plausibly see verification demand with no comparable public database currently serving them. No specific statistic, market size, or query volume is claimed anywhere below that can’t be traced to the source list itself — where this report can’t verify a number, it says so instead of guessing. See the Caveats section for the full discipline this was written under.

3. The coverage map

Four categories, well covered — and their frontiers.

“Covered well by” lists the databases from ASURIQ’s own methodology page. “Frontier gaps” are adjacent domains reasoned out from that same list, never invented from outside it.

Science & health
Covered well byPubMedClinicalTrials.govFDAWHOEurope PMCarXiv
Frontier gaps
Dietary supplements & nutraceuticals
FDA oversight of drugs — the pre-market efficacy review that makes the pharmaceutical coverage above so strong — doesn't extend the same way to supplements, which face no comparable pre-market efficacy requirement. A claim like “this supplement lowers cholesterol” sits one step outside that coverage: no public database indexes supplement-specific efficacy evidence at the same depth as PubMed and ClinicalTrials.gov index drug trials.
General consumer product safety
FDA covers drugs, devices, and food. Recalls and hazard reports for the much larger category of general consumer goods — toys, electronics, furniture — live in separate, agency-specific systems that aren't wired into the same claim-verification path as the free-tier sources above.
Complementary and alternative medicine
Claims about acupuncture, homeopathy, or chiropractic care are common in everyday conversation, but have far thinner representation in ClinicalTrials.gov and PubMed's indexing than pharmaceutical claims covering comparable ground.
Financial & regulatory
Covered well bySEC EDGARFREDBLSWorld Bank
Frontier gaps
Private-company financial claims
EDGAR's disclosure requirement is triggered by public trading. A claim about a private company's revenue, valuation, or funding round has no comparable public filing to check it against — the coverage gap is structural, not incidental.
Consumer financial product terms
FRED, BLS, and World Bank all operate at the aggregate economic level. A claim about one specific bank's interest rate or one insurer's premium doesn't have an institution-level public database at comparable resolution to what these sources provide for the economy as a whole.
Digital assets & cryptocurrency
Claims about token supply, protocol mechanics, or on-chain activity fall outside issuer disclosure (EDGAR), macro time series (FRED), and national accounts (World Bank) all at once — a genuinely uncovered category, not an edge case of an existing one.
Government & legislative
Covered well byCongress.govDOI Foundation
Frontier gaps
State and local legislative activity
Congress.gov indexes federal bills. Fifty state legislatures each run separate, non-uniform systems, so a claim about what a specific state bill actually says has no single comparable public database the way federal legislation does.
Municipal government records
City council actions, local ordinances, and municipal budget decisions are claims made constantly in local reporting, with no aggregated public index comparable to what exists at the federal level.
Sub-statutory agency guidance
The operational detail under a law — agency rules, guidance documents, interpretive memos — is more fragmented and harder to verify than the statute text itself, which is the level Congress.gov actually covers.
Reference & general
Covered well byWikipediaWikidataCrossRefOpenAlexGDELTGoogle Fact Check
Frontier gaps
Non-English-language claims
Corroboration density in Wikipedia, Wikidata, and GDELT skews toward English-language, Western-sourced coverage. Claims rooted primarily in non-English media have measurably thinner cross-referencing in these same sources — a coverage gradient within existing sources, not an absence of them.
Hyperlocal claims
GDELT is built for global event-level data. A claim about a specific local business, a neighborhood incident, or a school-district decision falls below the resolution these general reference sources operate at.
Social-platform origin & virality
Wikipedia and Wikidata are curated and lag real-time platform activity by design. No comparable general-reference source tracks where a claim or clip actually originated on a social platform.
4. What this means

The null map is market discovery.

None of the twelve gaps above are ASURIQ product announcements — they are observations about the information landscape, exactly as an auditor might note which company records exist and which don’t before anyone asks it to audit anything. “We cannot verify this from public sources” is a different statement from “this is unverifiable,” which is a different statement again from “this is false” — the same lexicon discipline that governs every ASURIQ verdict governs this report too. A reference database built to close any one of these gaps would be licensable from the day it exists, precisely because the demand for it is visible before the database is built. That is the finding: not that these domains are unimportant, but that nobody has yet built the public record they deserve.

5. Caveats

What this report is not.

This report does not claim to measure how often any of the domains above are actually asked about, does not estimate a market size in dollars, and does not assert that any named database is the only one that could ever exist for a domain. It is a reasoned, qualitative reading of ASURIQ’s own published coverage list, current as of the methodology version cited below, and will be revisited as coverage grows.

Methodology version referenced: v1.0 (effective 29 July 2026).
Permalink: https://asuriq.dev/reports/coverage-frontier-2026

Your data stays yours
Prompts stay with your provider. We see analysis metadata only.
Keys never stored
One-way hash for authentication. Your credentials pass through. Never persist.
~2 second responses
Single-model cognitive tools return in about 2 seconds.
37 of 100 flagged
We ran 100 ChatGPT answers through verification. See the study →

Read the standard this report reasons from.

Every category above traces to the same published methodology.