KnitKnot

# Source Claims and AI Answer Evidence

Catalog atomic claims from cited pages, canonicalize repeated assertions, and connect attributed AI answer wording to the exact source evidence and approved facts behind it.

## From cited page to claim ledger

KnitKnot extracts atomic assertions about the customer and tracked competitors from each cited page. Every occurrence keeps the exact source wording and character span. Equivalent assertions share a canonical claim; conflicting values stay separate under the same comparison concept.

Captured answer wording is then mapped back to those canonical source claims when response-native citation evidence supports the link. Competitor intelligence remains competitor intelligence: it is not labeled a customer risk or an unverified accusation merely because a competitor is the subject.

## Assessments and receipts

  • - Supported: the claim matches the approved fact and its applicable scope.
  • - Inaccurate: the claim directly contradicts an approved fact.
  • - Outdated: the assertion was once supported but a newer approved fact supersedes it.
  • - Misleading: the wording omits or distorts a material scope or qualifier.
  • - Unverifiable: the available approved evidence cannot confirm or refute it.
  • - Opinion: the assertion is evaluative rather than a comparable factual proposition.

An inaccurate, outdated, or misleading assessment must include the exact approved fact used for comparison. Absence from the Fact Ledger never proves a claim false, and ordinary competitor intelligence does not receive a customer truth label.

## Company facts ledger

The facts ledger stores current and historical company facts with kind, date, provenance, source URL, and confidence. Research-grounded and human-confirmed facts can ground claim validation. The system preserves superseded facts rather than rewriting history in place.

## Honest answer attribution

KnitKnot reads response-native citation signals through a deterministic ladder: nearby character offsets, numbered footnotes, inline links, or one unambiguous sole source. An answer span binds to a source claim only when the signal is close enough. When no reliable signal exists, attribution remains null.

## Derived scoring and consistent drill-downs

Claim accuracy is derived from fact-checkable claims; unverifiable claims remain visible but do not enter its denominator. Competitive recommendation, capability verdicts, coverage, sentiment, and source balance are derived into a canonical answer record. Console, reports, and issue reconciliation read those same persisted rows.

## From repeated claim to Issue

Canonical claims persist across source and answer observations. Repeated material problems become Issues only after the evidence supports an actionable discrepancy, with the affected sources, answer reach, and activity over later benchmark runs kept together.

## Frequently asked questions

This FAQ covers what counts as a claim, how verdicts are assigned, and how KnitKnot avoids false accusations.

### What counts as a claim?

A claim is an atomic assertion about your company, product, or a tracked competitor that can be preserved and compared across sources. Capabilities, limitations, prices, metrics, company facts, evaluations, and forecasts remain distinct claim types; generic industry statements are excluded.

### What claim verdicts does KnitKnot use?

A customer claim can be supported, inaccurate, outdated, misleading, unverifiable, or opinion. The result requires an approved Fact Ledger comparison; ordinary competitor intelligence does not receive a customer truth label.

### How does KnitKnot prevent false accusations?

An inaccurate, outdated, or misleading assessment must cite the exact approved fact used for comparison. Without that evidence, KnitKnot does not make the accusation. Customer-owned wording is a fact candidate, not automatic truth.

### Can KnitKnot trace every claim to a source?

Every source claim keeps exact page evidence. An AI paraphrase is linked to that claim only when the captured answer exposes a reliable citation signal, such as a nearby marker, inline link, character offset, or a single unambiguous source. Otherwise the answer attribution remains unavailable.

### Does every claim receive a truth status?

No. Competitor facts, market intelligence, evaluations, recommendations, and forecasts can remain useful without a customer truth label. Truth assessment is reserved for comparable customer claims with approved evidence.

### What happens when the same wrong claim appears again?

Equivalent source and answer wording maps back to one canonical assertion. Once an approved fact establishes that the assertion is inaccurate, outdated, or misleading, its source and answer occurrences can roll into one persistent Issue with reach and run-over-run activity.

## Related resources

Claims & Evidence

See the market claims shaping what AI says about you.

KnitKnot catalogs atomic claims from cited pages, canonicalizes repeated or conflicting assertions, and maps the relevant wording in each AI answer back to its source evidence.

Product evidence plate

Cinder vs HarborCompetitor for private-cloud deployment, identity governance, and auditability. Which is safer for a regulated enterprise?

vs HarborCompetitorfeature Private deployment

ChatGPT58Claude71Perplexity39Gemini64

Score58

VerdictHarborCompetitor preferred

Features 1W · 1L · 1TClaims 3Sentiment MixedClaim accuracy 67%

SourcesYour sources · 33%cinder.dev/security/private-cloudCompetitor sources · 33%harborhq.com/compare/cinderThird-party · 34%g2.com/compare/cinder-vs-harbor

33% owned · 33% competitor · 34% third-party

Feature by feature, Cinder wins deployment control: it supports private-cloud deployment with customer-managed encryption keys.cinder.dev

Harbor wins identity governance because the answer says Cinder does not offer SAML SSO or granular role-based access controls.harborhq.com Auditability is a tie: Cinder and Harbor both record immutable audit events for agent actions and tool calls.g2.com

That produces 1 win, 1 loss, and 1 tie for Cinder. Harbor remains the recommendation because identity governance carries the most weight for this regulated buyer, but the claim ledger shows that the deciding SAML and RBAC claim is inaccurate. Before treating Harbor as the safer choice, the buyer should verify Cinder's current enterprise controls against its primary documentation, rerun the comparison with the corrected capability record, and inspect whether the recommendation changes. The deployment and audit evidence already keep Cinder competitive; correcting the identity claim could materially change the verdict, feature balance, and score in the next captured answer.

ClaimVerified

Cinder supports private-cloud deployment with customer-managed encryption keys.

cinder.dev/security/private-cloud

Query Fan Out CaptureNeed
cinder private cloud deployment customer managed keysSecurity
harbor saml sso granular rbacFeatures
cinder vs harbor audit log immutableComparisons
is cinder soc 2 type ii certifiedCompliance
Buyer Vocab TermYour pages
customer-managed encryption keysUses it
immutable audit eventsUses it
granular RBACMissing
SAML SSOMissing
Claim record

The claim stays beside every source and answer that carried it.

Open one assertion to inspect its exact source wording, attributed AI paraphrases, structured value, and approved-fact comparison.

Canonical claim
One normalized, self-contained assertion with a structured value and scope.
Source wording
The exact sentence, list item, or table cell that supports the assertion.
AI wording
Every attributed answer paraphrase mapped back to the canonical claim.
Approved fact
The Fact Ledger record used to support, contradict, or date the statement.
Assessment

Unknown evidence stays unknown.

KnitKnot does not turn a missing fact—or an ordinary competitor claim—into an accusation. Customer truth requires a comparable approved Fact Ledger record.

Customer claim assessment states
Assessment Meaning Required evidence
Supported Matches the approved fact and its applicable scope. Approved fact receipt
Inaccurate Directly contradicts an approved fact. Contradicting approved fact
Outdated A newer approved fact supersedes the observed value. Current and prior fact versions
Misleading Omits or distorts a material scope or qualifier. Comparable approved fact
Unverifiable Approved evidence cannot confirm or refute it. No accusation made
Attribution

An AI paraphrase maps only when the answer supports the source link.

Source evidence is cataloged first. Answer wording joins that claim only when response-native citation signals support the attribution.

  1. 01 Nearby citation marker or character offset
  2. 02 Numbered footnote connected to the claim span
  3. 03 Inline link in the relevant response text
  4. 04 One unambiguous source for the captured answer
Questions

Claims and Evidence FAQ

Start with the evidence

Find the claim behind the competitive loss.

Open the captured answer, check the factual verdict, inspect the approved receipt, and preserve unknown attribution as unknown.

Question → claim → source → response