Compare · credit
AI Perception Signals vs Traditional Credit Bureau Data
Credit bureau data describes payment history and credit capacity. AI perception measures how models currently assess market standing, positioning, and risk language.
VeritasScore is not a credit score and does not predict probability of default. It measures AI-mediated reputation and positioning. Credit bureau data remains the foundation for capacity and history. The perception layer adds a forward-looking view of how models synthesize the public footprint — a view that is already influencing how names are framed inside institutions that use AI research.
Live report surface
VeritasScore · multi-model distribution
Northline Analytics · illustrative fixture on the real UI. Not a screenshot.
Score by LLM model
620Overall · 300–870
Side-by-side
Visibility
90
Brand Visibility component (max 150). Decomposable, monotonic.
Position
72
Market Position. Omission on category queries pulls this down.
Preference
22
Brand Preference from Focus Groups (max 40).
Cross-model gap
130 pts
ChatGPT 680 vs Claude/Grok 550. Near the paper’s 90th percentile (134).
Complementary use
The bureau file answers whether the borrower has the history and capacity the institution requires. The perception layer answers how large language models currently describe standing, positioning, and risk language — including whether that view is model-dependent. Used together they show lag between stated history and the narrative the market already holds. They should not be averaged into one number or treated as substitutes.
Monitoring implications
Four of five judge models are individually stable; split-perception for those models is a property of the model–evidence pair. The retrieval-grounded model can swing more than 200 points on a thin citation profile. Stable splits mean any single-model research note is tooling-dependent. High retrieval volatility is itself a flag that the public evidence base is thin.
Periodic re-assessment of portfolio names creates a light time series of perception state that can be reviewed alongside financial monitoring. Material movements, especially when accompanied by rising polarization or new high-severity objections, warrant attention even if financial metrics remain stable.
What a wide split means in a credit context
- The AI-assisted view of the borrower is highly model-dependent.
- The public evidence base is often thin or contested — itself a diligence-relevant fact.
- A single-assistant framing note is not a stable reading of market standing.
- Do not fold the spread into PD. File it beside the bureau pull as supplementary evidence.
Free analysis
Run it on any company
Public URL. Same funnel as the homepage. No card required.
Enter a public website
Multi-model · VeritasScore · ~5 minutes
See a sample dossier →FAQ
Short clarifications — positioning and proof live in the sections above.
Is VeritasScore a credit score?+
No. It does not predict probability of default. It measures AI-mediated reputation, positioning, and risk language, and is designed to sit beside bureau history, cash-flow analysis, and collateral assessment.
How should a wide split be read in a credit context?+
As a finding that the AI-assisted view of the borrower is highly model-dependent. It usually indicates a thin or contested public evidence base and supports additional human review of narrative strength and market positioning. Do not average it into PD.
Can this be used for portfolio monitoring?+
Yes. Periodic re-assessment produces a light time series of perception state that can be reviewed alongside conventional financial monitoring. Material movements, especially with rising polarization or new high-severity objections, warrant attention even if financial metrics remain stable.
Does high retrieval volatility mean the model is broken?+
Not under this reading. A retrieval-grounded model inherits the volatility of whatever it retrieves. Firms with thin citation profiles offer little for that retrieval to stabilize on. The volatility is a measurement of evidence thinness.
Where does this sit next to the bureau pull?+
The bureau remains primary for lending decisions. The perception layer is a supplementary signal for initial framing, underwriting discussion, and ongoing monitoring — making an already-active AI influence measurable rather than unrecorded.