◐MEDIUM CONFIDENCE

Onboarding implementation services are distinct from the subscription license

Both prongs of the distinct test were applied. Medium confidence reflects ambiguity in factor (b) — separately identifiable in the context of the contract — given the subscription cannot be used without successful onboarding.

Reasoning trace⬢ claude-opus-4-7 ⌥ prompt distinctness-saas-v1.4 ◆ self-consistency 0.84 · click ↓● RAG · 3 chunks · click ↓

Chain of thought

The customer can benefit from the onboarding services on their own — they receive a configured environment, migrated data, and SSO integration which have value independent of the ongoing subscription. Multiple third-party implementers exist in this market and the customer could reasonably engage one. The promise to deliver onboarding is separately identifiable from the subscription because the integration is generic (configuration of the customer's own resources, not customization of NorthStar's platform). Therefore both prongs are met and the onboarding is a distinct PO. Confidence is medium rather than high because this judgment hinges on factor (b) — a reviewer should confirm the integration depth from the SOW.

Retrieved authority · 3 chunks · cosine ≥ 0.83 · post-checked against verified-registry

CODIFICATION · ASC 606-10-25-19

A good or service that is promised to a customer is distinct if both of the following criteria are met: (a) The customer can benefit from the good or service either on its own or together with other resources that are readily available to the customer… (b) The entity's promise to transfer the good or service to the customer is separately identifiable from other promises in the contract…

CODIFICATION · ASC 606-10-25-21

Examples of factors that indicate that two or more promises to transfer goods or services to a customer are not separately identifiable include… the entity provides a significant service of integrating the good or service with other goods or services promised in the contract…

INVESTOR FRAMEWORK · Modernized investor framework · judgment screen · multi-deliverable check

Multiple-deliverable contracts trigger the platform's full distinctness analysis under ASC 606-10-25-19 and -25-21. (Modernized investor framework · judgment-quality screen.)

Structured output

PerformanceObligation{
  po_id:               "po-2-onboarding",
  distinct_capable:    true,
  distinct_in_context: true,
  overall_distinct:    true,
  authority_cited:     ["606-10-25-19", "606-10-25-21", "Modernized investor framework · judgment screen · multi-deliverable check"],
  confidence:          "medium",
  consistency_score:   0.84,  // 5 samples; 4/5 on factor (b)
  requires_human_review: true,
  model:               "claude-opus-4-7",
  prompt_v:            "distinctness-saas-v1.4",
  judgment_id:         "018f2a4b7c80e1faa910d4c3de0f5b77"
}

◆ MethodologyHow we arrived at confidence = medium◐CISC 0.84

We use Confidence-Informed Self-Consistency (CISC, arxiv 2502.06233): the same prompt is sampled 5 times at temperature 0.3; each sample produces an independent judgment. We then measure two agreement metrics — direction agreement (does every sample reach the same conclusion?) and reasoning agreement (do they reach it for the same reason?). The weighted vote becomes the score.

The 5 sampled judgments

#	Direction	Factor (a) · capable of being distinct	Factor (b) · separately identifiable reasoning	Cluster
1	distinct = true	✓ standalone benefit	third-party implementers exist	✓DOMINANT
2	distinct = true	✓ standalone benefit	configuration not customization	◐MINORITY
3	distinct = true	✓ standalone benefit	third-party implementers exist	✓DOMINANT
4	distinct = true	✓ standalone benefit	subscription is functional standalone	◐MINORITY
5	distinct = true	✓ standalone benefit	third-party implementers exist	✓DOMINANT

CISC weighted score · how 0.84 was computed

direction_agreement = 5/5 = 1.00

reasoning_agreement = 3/5 (3 samples cite "third-party implementers") = 0.60

cisc_score = 0.6 × direction + 0.4 × reasoning

= 0.6 × 1.00 + 0.4 × 0.60 = 0.84

Confidence tier mapping

✓HIGH ≥ 0.95

auto-applied

◐MEDIUM 0.80 – 0.95

human review suggested

⚠LOW < 0.80

hard-blocked · human required

Bottom line: the AI is unanimous that PO-2 is distinct (100% direction agreement). Where it splits is why — 60% of samples emphasize that third-party implementers exist; 40% emphasize technical genericness. Both are valid factor-(b) supports under ASC 606-10-25-21, but a reviewer should pick the dominant rationale before audit.

Modernized investor framework Q1 — judgment-laden recognition areas. Where in this company's revenue policy is the recognition judgment most concentrated? For NorthStar, the answer is the PO-2 onboarding distinctness call. The investor analytical follow-up requires documenting the distinctness reasoning. Recommend confirming integration depth from the SOW before sign-off.Modernized investor framework · Q1 · ASC 606-10-25-19 / -25-21