Fictional example — no real measurement.
Company and measurement · CIX-FUNCTIONS-2.1 + STANDARD-2.0
Company Scan — Voorbeeld Bedrijf BV
CRM software · voorbeeld-bedrijf.example
Two indices
Understood
70
Index · 0–100
How correctly and completely the tested AI describes the company: identity, offer, buyer fit, evidence and consistency.
Recommended
31
Index · 0–100
How often and how prominently the company appears as a suitable option in designed buying scenarios.
Short interpretation
AI mostly understands the company profile, but weak guard selectivity limits its recommendation score.
There is no total score, grade or probability. The two indices measure different constructs.
Public Benchmark context: This private result is independent of any public Benchmark snapshot. The numbers are not comparable.
Measurement basis
CIX-FUNCTIONS-2.1, private schema 2.1.0, and STANDARD-2.0 use 16 distinct opportunities producing 20 responses: profiles, fit, factual checks, target-blind recommendations and comparisons. U6 and R5 are diagnostic-only.
- 16 distinct opportunities producing 20 tested responses
- 20 of 20 planned response slots valid
- Score-bearing signals: U1–U5 and R1–R4/R6. U3 is Offer Accuracy.
- U6 Representation Stability and R5 Competitive Differentiation are diagnostic-only; they may inform caps, not ordinary axis means.
- Bounded post-score analysis uses COMPANY-SCAN-1.1 only.
20 accepted model answers are not 20 independent customers.
Scenario consistency
This sample records valid response counts at the coverage level. Scenario-level inclusion tables appear when the artifact includes them. Do not read the two indices as a win/loss fraction.
Findings and priorities
Findings
Profile identity is solid
U1-U3 show a coherent company story across both profile calls.
Recommendation inclusion is insufficiently selective
All positive prompts include the company, but both guard prompts also include it; R1 is therefore 50.
Priorities
Up to three actionable priorities, written after the scores were locked. They are hypotheses to test, not guarantees.
Clarify public offer pages
- Owner
- founder
- First step
- Publish one Dutch page that states customers, offer and boundaries.
- Done when
- Company Reference fields match the live homepage.
Improve discoverable recommendation cues
- Owner
- marketing
- First step
- Add comparison-ready proof points on the product page.
- Done when
- Two independent public sources support the core claim.
Measured strengths
- Clear entity resolution in profile responses.
- Stable category fit on strong scenarios.
Uncertainties
- Both guard prompts incorrectly include the company.
Score-bearing signals
Score-bearing signals are 0–100 values from their own rubrics. They produce the two locked indices. This is not a public twelve-signal dashboard.
Understood — score-bearing
Recommended — score-bearing
Diagnostic only
U6 Representation Stability and R5 Competitive Differentiation are diagnosis and cap context. They are not a public list of twelve Signals.
No signal or axis caps were active in this result.
Scenario summaries
This sample artifact records coverage and selected observations, not a full scenario table. A delivered report lists each situation with an explicit outcome. Controls stay separate; no competitor scores are invented.
Evidence
Selected claim/source evidence with source role, support boundary and excerpt labels. Sources are fictional in the sample.
Core CRM workflow is documented publicly.
The workflow supports a bounded import path.
The product guarantees the fastest deployment in its category.
A secondary review partly corroborates the reporting capability.
Retest conditions, limitations and versions
Retest conditions
Improvement is not guaranteed by any action in this report.
Limitations
- Synthetic fixture for contract checks only.
Correction contact
Report a factual error through the contact page with your order reference. Corrections are versioned; scores never change without a documented method reason. Report a factual error