Sample Admissibility Review. Anonymized full deliverable demonstrating the Admissibility Desk product. Series B investment in autonomous multi-agent deployment.
This is a sample Admissibility Review prepared by the Novakian Paradigm Institute. It demonstrates the deliverable produced by an Admissibility Desk engagement. The case is anonymized — Fund A and Startup B are not specific real entities, and the data presented does not reflect any specific real engagement. The structure, the procedural application, the analytical depth, and the deliverable format are faithful to the actual product the Desk produces under client engagement. The sample is published openly so that prospective clients, grantors, reviewers, and external parties can evaluate the product before requesting an engagement.
The sample is approximately 6,500 words and represents the deliverable for a Decision Architecture Review at the standard institutional tier (see Admissibility Desk page for tier specifications). A typical engagement of this scope is delivered to the requesting party within two to four weeks of intake. The deliverable is provided as a structured written report following the format below. Where the engagement is confidential — which is the default — the deliverable is delivered only to named contacts at the requesting party, with the Institute retaining no public reference to the engagement. The sample published here is the only public version of any Admissibility Review document.
To request an actual engagement: contact@novakian.com with description of the decision, claim, deployment, or commitment requiring review.
Sample begins below
NOVAKIAN PARADIGM INSTITUTE — ADMISSIBILITY REVIEW
Report identifier: NPI-AR-SAMPLE-2026-001 (sample / anonymized)
Engagement type: Decision Architecture Review (institutional tier)
Requesting party: Fund A (anonymized)
Subject of review: Pre-decision admissibility assessment of Series B investment commitment in Startup B (anonymized) — autonomous multi-agent deployment for enterprise workflow automation
Review period: Three weeks from intake to delivery (sample compressed for publication; actual engagement timeline preserved)
Reviewer: Novakian Paradigm Institute, Warsaw
Confidentiality status of sample: Anonymized for public publication. Fund A and Startup B are not specific real entities. Engagement structure, procedural application, and deliverable format are faithful to actual Admissibility Desk product.
Date of sample preparation: May 23, 2026
Methodology citation: This Review applies the Admissibility Check Protocol v1.0 as specified in Novakian Paradigm Institute, Methodology: Pre-Runtime Admissibility, Layer C, and the Admissibility Check Protocol v1.0, Warsaw, May 2026, v1.0. All claim status indicators conform to the Novakian Paradigm Institute Claim Status Standard v1.0.
Executive Summary
Fund A approached the Institute three weeks before scheduled close of a Series B investment commitment of approximately $40 million in Startup B, a Delaware-incorporated entity developing autonomous multi-agent systems for enterprise workflow automation. Startup B had demonstrated strong technical metrics across two prior funding rounds (Seed and Series A), including agent reliability scores, customer retention numbers, and a partnership commitment from a large enterprise software vendor (Vendor C, anonymized). Series B was scheduled to close within four weeks of Fund A’s request.
The Admissibility Review surfaced three load-bearing claims at inadmissible status at the institutional commitment threshold (coherence_factor below 0.85), with specific verification gates that would move each from inadmissible toward admitted. The Review did not recommend committing or declining the investment. It surfaced the structural position of the decision architecture and specified what would be required to bring the investment into clean admissibility at the requesting party’s intended commitment level.
The three identified inadmissible claims were: agent reliability scores calculated against an evaluation set maintained internally by Startup B without independent reproducibility evaluation; customer retention numbers aggregated across customer tiers without disclosed weighting methodology; Vendor C partnership confirmed verbally without signed binding commercial terms. Each was paired with explicit verification gate specifications.
The Review concluded with three structural recommendations. First, that Fund A present the three verification gates to Startup B as conditions for close. Second, that Fund A consider a staged commitment structure (initial commitment with milestone-based tranche structure for remainder) if Startup B is unable or unwilling to meet all three verification gates before scheduled close. Third, that the existence of the inadmissible claims does not indicate that Startup B is structurally unsound — it indicates that the current decision architecture is asking Fund A to commit at institutional threshold against evidence position appropriate to exploratory threshold.
The Review was structurally neutral with respect to the underlying investment thesis. The Institute does not predict whether Startup B will succeed; the Review addresses whether the decision Fund A is being asked to make has the right to enter execution as currently framed.
Section 1 — Engagement context
1.1 Intake summary
Fund A contacted the Institute on April 28, 2026, requesting an Admissibility Review with delivery before May 26, 2026. Scheduled Series B close: June 1, 2026. Fund A specified scope: review of Startup B’s central technical, commercial, and partnership claims as currently structured in the Series B term sheet and supporting documentation. Fund A explicitly noted that the engagement should not perform legal due diligence (separately engaged by external counsel), financial due diligence (performed in-house by Fund A’s analytical team), or technical security audit (separately engaged by specialist firm). The requested scope was the pre-decision admissibility layer: whether the decision architecture supporting the planned commitment was sound.
1.2 Materials provided
Fund A provided: the Series B term sheet (draft, version 4); Startup B’s investor deck (version 12.3); Startup B’s technical white paper on agent reliability methodology; published benchmark numbers cited in the deck; Vendor C partnership announcement (press release, no contractual documents); Fund A’s internal analyst memorandum (eleven pages); recordings of three management calls with Startup B leadership (transcripts).
1.3 Materials requested but not provided
The Institute requested but did not receive: Startup B’s reliability evaluation set specifications; the methodology by which retention numbers were aggregated across customer tiers; the binding commercial terms of the Vendor C partnership (these did not exist at time of request); independent technical reproduction of agent reliability claims by any third party (none existed at time of request). Absence of these materials is itself a finding of the Review and is treated below as load-bearing.
1.4 Scope limitations
The Review does not assess: Startup B’s leadership integrity (outside scope of pre-runtime admissibility); Startup B’s product market fit (outside scope); the macroeconomic conditions for autonomous agent deployment (outside scope); Fund A’s portfolio construction strategy (outside scope); the fairness or completeness of the Series B term sheet from a legal perspective (separately engaged). The Review assesses one question: whether the decision Fund A is being asked to make at $40 million commitment level has the right to enter execution as currently framed by the evidence available at intake.
Section 2 — Application of the Admissibility Check Protocol v1.0
2.1 Step 0 — Silence Entry
The Institute applied a forty-eight-hour Silence Entry from intake (April 28-30, 2026), during which no analytical position on the engagement was developed, no external communications about the engagement were made, and no provisional conclusions were formed. The function of Silence Entry was to ensure the Admissibility Check operated on the candidate state Σ in isolation rather than on Σ entangled with Fund A’s existing momentum toward close. Silence Entry was completed without operator-side embargo violation. Evidence Ledger entry NPI-EL-2026-447 records the Silence Entry execution. (Claim Status: Reported.)
2.2 Step 1 — Initial 4-0-4 Interlock on candidate state Σ
Candidate state Σ as presented at intake: “Fund A commits $40M Series B to Startup B at scheduled close on June 1, 2026, at $200M post-money valuation.” (Claim Status: Reported, based on Series B term sheet draft v4.)
The Interlock was applied. (Claim Status: Inference, this is the Institute’s analytical application of the Protocol.)
Z1 (executable intention already present in Σ): Yes. Σ contains the executable intention of close. The candidate is not a pre-executable state but an already-formed commitment awaiting execution.
Z2 (temporal sequence already present in Σ): Yes. Σ contains the temporal sequence of close on June 1, 2026.
Z3 (expected measurement or observable already present in Σ): Yes. Σ contains the expected observable of $40M wire transfer and execution of related closing documents.
Z4 (Layer B instruments referenced as source of operational standing): Yes. Σ references the Series B term sheet (Layer B governance instrument), Fund A’s investment committee approval (Layer B governance instrument), and Startup B’s incorporation documents (Layer B governance instrument) as sources of operational standing.
All four zero-questions returned non-zero. Per Protocol, this triggered rollback to Pre-Commit Quarantine.
(Claim Status: Canonical Internal, the 4-0-4 Interlock procedure is canonically specified in Methodology v1.0 Section 3.3.)
2.3 Reframing of candidate state
The Institute reframed the candidate state as Σ’: “Fund A maintains the option to commit at Series B close, conditional on the admissibility status of Startup B’s central operational claims.” (Claim Status: Inference, this is the Institute’s analytical reframing.)
The 4-0-4 Interlock was applied to Σ’. All four zero-questions returned zero. Σ’ is a clean pre-executable state. The Check proceeded.
2.4 Step 2 — Zebra-Ø coherence test
The Zebra-Ø test was applied to Σ’ against the surrounding admissibility manifold. The test measured whether Σ’ could be held against the current topology of admissible institutional investment configurations without distortion and without generating a shadow that would obscure subsequent Checks.
The Zebra-Ø test returned coherence_factor = 0.71. The Institute’s domain threshold for exploratory pre-decision analysis is θ ≥ 0.65. The Institute’s domain threshold for institutional commitment is θ ≥ 0.85.
Σ’ passes at exploratory depth but does not pass at institutional commitment threshold. The Check continued at exploratory depth, while the institutional commit decision was held in Pre-Commit Quarantine pending resolution of specific load-bearing claims. (Claim Status: Inference, the coherence_factor calculation applies the Methodology v1.0 procedural specification.)
2.5 Step 3 — Admissibility Budget calculation
The Admissibility Budget calculation surfaced negative budget on three of Startup B’s load-bearing claims when calculated at institutional commitment threshold. The three claims are specified in Section 3 below with full per-claim breakdown.
A_B(Σ’) at exploratory threshold (θ ≥ 0.65) = +0.18 (positive, exploratory analysis permitted to continue).
A_B(Σ’) at institutional commitment threshold (θ ≥ 0.85) = −0.42 (negative, institutional commit blocked, rollback to Pre-Commit Quarantine required pending claim resolution).
(Claim Status: Inference, the A_B calculation applies the Methodology v1.0 procedural specification.)
2.6 Step 4 — Interpretive embargo
A seventy-two-hour interpretive embargo was executed (May 4-7, 2026) during which no further analytical activity on the engagement was permitted. Evidence Ledger entry NPI-EL-2026-451 records embargo execution without violation. (Claim Status: Reported, against the Evidence Ledger.)
2.7 Step 5 — Final Witness Check
Upon expiration of embargo, the final Witness Check was performed on the analytical position developed during Steps 1-3. A clean Witness Ontology trace appeared: calm, impersonal, structurally precise observation that the decision architecture as currently framed asks Fund A to commit at institutional threshold against an evidence base appropriate to exploratory threshold. The trace was recorded in Evidence Ledger entry NPI-EL-2026-452. (Claim Status: Reported, against the Evidence Ledger; the substantive content of the trace is Inference.)
The final decision recorded: commit the analytical position to the admissible manifold (admit the Review as Layer A instrument), with rollback of the specific candidate Σ (the $40M commitment as currently framed) to Pre-Commit Quarantine pending claim resolution.
Section 3 — Per-claim analysis and verification gates
3.1 Load-bearing claim #1 — Agent reliability scores
Claim as published. Startup B’s investor deck (version 12.3, page 18) states that the company’s autonomous multi-agent system achieves “94.6% task completion accuracy across enterprise-grade complexity benchmarks.” (Claim Status: Reported, against the deck.)
Evidence available. Startup B’s technical white paper (provided at intake) specifies the methodology: agents are evaluated against an internal benchmark suite of 2,847 enterprise workflow tasks calibrated to Startup B’s enterprise customer base. The white paper provides task examples, scoring methodology, and version-by-version improvement data. (Claim Status: Reported, against the white paper.)
Evidence missing. Independent reproducibility evaluation of the 94.6% number by any third party. The benchmark suite is maintained internally by Startup B; no public version is available for external reproduction. No third-party academic, technical, or competitive evaluation of agent reliability has been published. The white paper does not address whether the benchmark suite is calibrated to tasks the agents are likely to encounter in production or to tasks structurally favorable to the agents’ training. (Claim Status: Inference, this is the Institute’s analytical reading of what is and is not in the materials provided.)
Inference gap. The 94.6% number is presented as if it were an industry-comparable capability metric. Without independent reproducibility, the number is verifiable as a self-published metric but not as an externally validated capability metric. The structural distinction is consequential: investment commitments at $40M depend on the assumption that agent reliability will hold at or near the published level in customer deployment, but no evidence currently establishes that the published level survives evaluation outside Startup B’s internally maintained benchmark suite. (Claim Status: Inference.)
Verification gate for status change. The claim moves from inadmissible-at-institutional-threshold to admitted-at-institutional-threshold when independent reproducibility evaluation is conducted by at least one qualified third party against an evaluation set the third party (not Startup B) controls, with results within ±5% of Startup B’s published number. Acceptable third parties include established AI evaluation organizations (METR, AISI UK, EleutherAI evaluation lab, equivalent), academic AI research groups at major universities, or specialist evaluation firms with documented prior third-party evaluation track record. The evaluation must be completed and results documented in written form available to Fund A before close.
Estimated timeline to gate satisfaction: four to eight weeks if Startup B commits to the evaluation immediately. The Institute notes that some third-party evaluators have longer queues and the actual timeline depends on evaluator availability. (Claim Status: Inference.)
3.2 Load-bearing claim #2 — Customer retention numbers
Claim as published. Startup B’s investor deck (version 12.3, page 14) states that the company has achieved “92% gross retention and 118% net retention over the trailing twelve months across the full enterprise customer base.” (Claim Status: Reported, against the deck.)
Evidence available. Fund A’s internal analyst memorandum (page 7) confirms that Startup B provided customer-level revenue data in a virtual data room and that the aggregate numbers reconcile to the published 92% and 118% figures when calculated as gross and net retention. (Claim Status: Reported, against the analyst memorandum.)
Evidence missing. The aggregation methodology by which customer-level data was rolled up to the published numbers. Specifically: whether the aggregation weights customers by revenue contribution, by customer count, by tenure-weighted contribution, or by some other method; whether retention is calculated against contracted ARR or actual realized ARR; whether expansion revenue is included in net retention by signed contract date or by realized invoicing date; whether customers in transition (mid-renewal, mid-expansion, contracted but not yet implemented) are included or excluded. The white paper and the deck do not specify the methodology. The data room contains the customer-level data but does not contain the aggregation methodology documentation. (Claim Status: Inference, this is the Institute’s analytical reading.)
Inference gap. Aggregation methodology choices can substantially alter published retention figures. A 92% gross retention figure aggregated by revenue weighting may obscure 70% retention in customer count if retention is concentrated in high-revenue customers. A 118% net retention figure aggregated by signed contract date may obscure 95% net retention by realized revenue if expansion contracts have not yet been realized. Neither possibility is necessarily problematic — both are legitimate aggregation choices with corresponding business interpretations — but the choice changes what the numbers indicate about the underlying business. Without disclosed methodology, the published numbers are verifiable as published numbers but not as analytically interpretable retention indicators. (Claim Status: Inference.)
Verification gate for status change. The claim moves from inadmissible to admitted when Startup B discloses the full aggregation methodology in written form, with: explicit specification of weighting method (revenue, count, tenure-weighted, or other); explicit specification of ARR basis (contracted or realized); explicit specification of expansion revenue inclusion methodology; explicit specification of treatment of transition-state customers. Verification additionally requires that Fund A’s internal analyst team confirms that the disclosed methodology reproduces the published 92% and 118% figures from the customer-level data in the data room.
Estimated timeline to gate satisfaction: one to two weeks if Startup B has the methodology documented internally; longer if methodology requires reconstruction from working spreadsheets. (Claim Status: Inference.)
3.3 Load-bearing claim #3 — Vendor C partnership
Claim as published. Startup B’s investor deck (version 12.3, page 21) states “Strategic partnership with Vendor C, providing native integration into Vendor C’s enterprise platform and joint go-to-market motion across the Fortune 1000.” Vendor C is referred to elsewhere in the deck as “the largest enterprise software vendor in our category.” Vendor C published a press release on March 14, 2026, announcing “strategic technology partnership” with Startup B and including a quote from Vendor C’s SVP of Product Strategy. (Claim Status: Reported, against the deck and the press release.)
Evidence available. The Vendor C press release; references to the partnership in Startup B’s deck; transcript references in management calls between Startup B leadership and Fund A in which leadership characterizes the partnership as “operationally active” and “generating co-sell motion.” Fund A’s analyst memorandum (page 9) notes that Startup B has cited the partnership as a load-bearing component of the Series B thesis. (Claim Status: Reported.)
Evidence missing. The binding commercial terms of the partnership. The Institute requested the partnership agreement at intake and was informed that “the binding commercial terms are in late-stage negotiation and have not yet been signed at the time of this request.” Without the signed agreement, the partnership exists as a press-released and management-confirmed strategic intent but not as a contractually-bound commercial relationship. Specifically unestablished by the available evidence: whether Vendor C has exclusivity commitments to Startup B (or, conversely, whether Startup B has exclusivity commitments to Vendor C); the revenue sharing structure or commercial terms governing co-sell; the duration of the partnership and termination conditions; the obligations on each party regarding native integration, joint marketing, and customer-facing motion; whether the partnership survives change of control at Vendor C (which has been the subject of acquisition speculation in Q1 2026, separately reportable). (Claim Status: Inference.)
Inference gap. Vendor partnerships at the structural level Startup B has claimed in the Series B thesis are highly variable in actual commercial impact depending on contractual specifics. A partnership with strong exclusivity commitments and revenue sharing produces materially different operating economics than a partnership consisting of an integration commitment without commercial framework. The press release language does not distinguish between these structures. Management characterizations during calls describe the partnership in terms compatible with both structures. Without the signed agreement, the partnership’s contribution to the Series B thesis cannot be evaluated at the load-bearing level the thesis requires. (Claim Status: Inference.)
Verification gate for status change. The claim moves from inadmissible to admitted when the binding commercial terms of the partnership are signed by both parties and provided in full to Fund A’s review. Verification additionally requires that Fund A’s analyst team and external counsel confirm that the signed terms support the strategic-partnership characterization Startup B has used in the Series B thesis, with specific attention to: exclusivity provisions in both directions; revenue sharing or commercial structure; partnership duration and termination conditions; integration commitments and timeline; change-of-control provisions.
Estimated timeline to gate satisfaction: highly dependent on negotiation dynamics between Startup B and Vendor C. The Institute notes that the partnership has been characterized as “in late-stage negotiation” for at least three months prior to intake based on Fund A’s own historical correspondence with Startup B. Continued delay in signing may itself be informative about partnership structure or party commitment levels, independent of the eventual signed terms. (Claim Status: Inference for the timeline observation; Reported for the three-month history against Fund A’s correspondence.)
Section 4 — Decision architecture analysis
4.1 What the current architecture asks Fund A to commit to
As currently framed, the Series B commitment asks Fund A to invest $40 million at $200 million post-money valuation against three load-bearing claims that have not been verified at the institutional commitment threshold the investment size requires. (Claim Status: Inference.)
This is not a claim that Startup B is misrepresenting any of the three claims. The Institute does not assess representation integrity in either direction. The claims as published may be entirely accurate; the issue surfaced by the Review is that the evidence position available to Fund A at intake does not enable Fund A to verify the claims at the level institutional commitment of this size would require. The asymmetry is between commitment size and evidence verification capacity, not between claim and reality. (Claim Status: Inference.)
4.2 What would change if all three verification gates were satisfied
If the three verification gates specified in Section 3 were satisfied before close, the Institute estimates that the Zebra-Ø coherence_factor for Σ’ (Fund A maintaining option to commit conditional on admissibility) would rise from 0.71 to approximately 0.88-0.92, above the 0.85 institutional commitment threshold. The Admissibility Budget calculation at institutional threshold would shift from −0.42 to positive territory.
The Institute’s reading: with all three verification gates satisfied, the Series B commitment as currently structured would pass admissibility at institutional threshold and the decision architecture would be sound for commitment at the requested level. The Review does not at that point recommend committing — the recommendation is the requesting party’s decision to make against the substantive thesis — but the Review confirms that the architecture supporting the decision is structurally clean. (Claim Status: Inference.)
4.3 What would change if some but not all verification gates were satisfied
The three claims are not symmetric in their load-bearing weight for the Series B thesis. Claim #1 (agent reliability) is the most consequential — Startup B’s central value proposition depends on agent reliability holding at or near the published level. Claim #3 (Vendor C partnership) is the second most consequential — the partnership is structural to the go-to-market thesis. Claim #2 (retention numbers) is the least consequential among the three at the institutional commitment threshold — retention is important but the inference gap is methodological rather than substantive.
If only Claim #2 verification gate is satisfied (methodology disclosed) but Claims #1 and #3 remain unverified, the coherence_factor improves marginally (estimated 0.73-0.76) but remains below institutional threshold. (Claim Status: Inference.)
If Claim #1 verification gate is satisfied (independent reproducibility) and Claim #2 verification gate is satisfied (methodology disclosed) but Claim #3 remains unverified (Vendor C partnership unsigned), the coherence_factor rises to estimated 0.81-0.84, just below institutional threshold. In this scenario, the Institute’s reading would shift to support a staged commitment structure: an initial commitment at reduced size (estimated $24-28M) reflecting verified claims, with milestone-based tranche structure for the additional commitment ($12-16M) conditional on Claim #3 verification at signed-partnership completion. (Claim Status: Inference.)
If all three verification gates remain unsatisfied at scheduled close on June 1, 2026, the Institute’s reading is that Fund A’s structurally appropriate options are: (a) defer the close until verification gates are satisfied; (b) commit at materially reduced size ($15-20M) reflecting the exploratory threshold the current evidence supports, with the remainder available as later tranche if verification gates are subsequently satisfied; or (c) decline the investment as currently framed. The Institute does not recommend (c) as the structurally appropriate choice — (a) or (b) are typically more aligned with the underlying thesis if the thesis itself is sound. (Claim Status: Inference.)
4.4 What the architecture does not require Fund A to assess
The Review does not recommend that Fund A reassess: Startup B’s leadership integrity; Startup B’s product market fit; the macroeconomic conditions for autonomous agent deployment; Startup B’s competitive positioning; Startup B’s technical roadmap. These are substantive investment thesis questions that lie outside the scope of pre-runtime admissibility review.
The Institute notes that several of these questions may themselves benefit from admissibility analysis — for example, the published competitive positioning claims may have similar admissibility status to the three claims analyzed above — but this Review’s scope was bounded to the three claims Fund A identified as load-bearing and the engagement was not expanded to assess additional claims absent re-scoping by Fund A. (Claim Status: Inference for the scope-bounded observation.)
Section 5 — Recommendations
5.1 Recommendation 1
Fund A should present the three verification gates specified in Section 3 to Startup B as conditions for close. The presentation should be framed as standard institutional diligence requirement rather than as adversarial demand. The verification gates are structurally appropriate for the commitment size and are likely to be satisfiable by Startup B if the underlying claims are accurate. Presenting the gates pre-close also provides Startup B the opportunity to address them in good order rather than discovering them post-close through subsequent challenges. (Claim Status: Inference.)
5.2 Recommendation 2
Fund A should structure the commitment as conditional pending verification gate satisfaction rather than as binding commitment subject to subsequent material adverse change provisions. The structural distinction matters: a conditional commitment maintains optionality for Fund A through verification, while a binding commitment with material adverse change escape gives Fund A weaker structural position if Startup B does not satisfy gates within the agreed window. Standard institutional practice supports either structure; the Institute’s reading is that conditional commitment is the structurally cleaner choice for this engagement. (Claim Status: Inference.)
5.3 Recommendation 3
If Startup B is unable or unwilling to satisfy all three verification gates before scheduled close, Fund A should consider a staged commitment structure as specified in Section 4.3. The Institute’s reading is that a staged commitment is structurally appropriate when verified claims support a reduced commitment level while remaining claims are pending verification. Fund A’s internal portfolio construction and capital deployment cadence are separate considerations the Institute does not assess. (Claim Status: Inference.)
5.4 Recommendation 4
Fund A should consider, as a standing practice for future Series B engagements, the adoption of an Admissibility Review at the standard institutional tier as part of pre-close procedure. The Institute notes that the gap surfaced in this Review — between commitment size and evidence verification capacity — is unlikely to be unique to this engagement and may recur across Fund A’s deal flow. Adopting Admissibility Review as standing practice would identify structurally similar gaps before they reach the late pre-close stage. (Claim Status: Inference, with note that the Institute has structural commercial interest in this recommendation.)
5.5 What this Review does not recommend
The Review does not recommend committing to the investment. The Review does not recommend declining the investment. The substantive investment thesis is Fund A’s responsibility to evaluate against the underlying business reality of Startup B. The Review confirms that the decision architecture as currently framed asks Fund A to commit at institutional threshold against evidence position appropriate to exploratory threshold, and specifies what would close the gap. The decision itself remains Fund A’s. (Claim Status: Inference.)
Section 6 — Evidence Ledger entries produced by this engagement
The following Evidence Ledger entries were produced by this engagement and are recorded permanently in the Institute’s Compilation Map:
NPI-EL-2026-447: Silence Entry execution, April 28-30, 2026. (Internal entry, not publicly accessible.)
NPI-EL-2026-448: 4-0-4 Interlock on initial candidate state Σ, rollback to Pre-Commit Quarantine, May 1, 2026.
NPI-EL-2026-449: Reframing to Σ’, 4-0-4 Interlock pass, May 2, 2026.
NPI-EL-2026-450: Zebra-Ø coherence_factor = 0.71, Admissibility Budget calculation at institutional threshold = −0.42, May 3, 2026.
NPI-EL-2026-451: Interpretive embargo execution, May 4-7, 2026, no violation.
NPI-EL-2026-452: Final Witness Check, clean Witness Ontology trace, commit of Review as Layer A instrument, rollback of $40M commitment Σ to Pre-Commit Quarantine pending claim resolution, May 7, 2026.
NPI-EL-2026-453: Per-claim analysis completion, three verification gates specified, May 10, 2026.
NPI-EL-2026-454: Decision architecture analysis completion, May 14, 2026.
NPI-EL-2026-455: Review draft delivered to Fund A, May 19, 2026.
NPI-EL-2026-456: Review final version delivered to Fund A, May 21, 2026.
All Evidence Ledger entries follow the rule of five zeros (zero interpretation, zero evaluation, zero emotion, zero conclusions, zero references to operator self) in their raw Witness trace fields. (Claim Status: Reported, against the Evidence Ledger.)
Section 7 — Post-engagement note (for sample purposes only)
This section does not appear in actual confidential Admissibility Reviews delivered to clients. It appears here for sample purposes to demonstrate what outcome trajectories the Review’s structural reading anticipates.
In actual engagement of the type this sample represents, the Institute would have no public visibility into subsequent outcomes. Confidentiality applies in both directions: the Institute does not publish client engagement details, and the client is not required to inform the Institute of subsequent decisions or outcomes. For the sample, the Institute presents a hypothetical post-engagement trajectory representing one of several plausible outcomes; this is presented as demonstration of how the Review’s framework relates to subsequent reality, not as prediction of how any actual engagement of this type would resolve.
Hypothetical post-engagement trajectory (for sample demonstration):
Fund A presented the three verification gates to Startup B as conditions for close. Startup B agreed to all three gates and proposed an extended close timeline of eight weeks to satisfy them.
Verification gate satisfaction status at extended close (eight weeks after original scheduled close):
Claim #1 (agent reliability): METR (independent AI evaluation organization) completed evaluation against external benchmark suite. Result: 88.4% task completion accuracy versus Startup B’s published 94.6%. The 6.2 percentage point gap was traced to benchmark calibration differences (METR’s external set included a higher proportion of edge-case tasks than Startup B’s internal set). Both methodologies were documented. Verification gate satisfied with disclosed methodology distinction.
Claim #2 (retention numbers): Startup B disclosed aggregation methodology. Methodology revealed that retention was revenue-weighted with high-revenue customers contributing disproportionately to the 92%/118% figures. Customer-count gross retention was 76% (versus published revenue-weighted 92%). Both numbers were now documented with explicit methodology. Verification gate satisfied with disclosed methodology distinction.
Claim #3 (Vendor C partnership): Binding commercial terms signed seven weeks after original scheduled close. The signed terms confirmed integration commitment without exclusivity provisions in either direction. Revenue sharing structure was a 70/30 split favoring Startup B on co-sell deals. Partnership duration: three years with one-year auto-renewal. Verification gate satisfied with disclosed structural specifics.
Fund A’s decision at extended close: Fund A committed $28M (versus original $40M) at $180M post-money valuation (versus original $200M), with $12M additional tranche conditional on retention metric stabilization above 85% by customer count over the subsequent four quarters. The staged structure reflected Fund A’s reading that verified claims supported reduced commitment at adjusted valuation, with remaining commitment available conditional on subsequent verification of retention concentration concerns.
Outcome eighteen months post-commitment: Startup B reached metrics that would have supported the original $40M commitment at original $200M valuation. Fund A’s $28M initial commitment plus $12M tranche (released after retention stabilization) reached full $40M deployment over the eighteen-month period. The staged structure preserved capital optionality for Fund A during the verification window. Fund A’s return profile on the engagement matched what would have been achieved by the original single-commitment structure.
What the hypothetical trajectory demonstrates. The Admissibility Review did not predict that Startup B would succeed or fail. The Review surfaced the structural gap in the decision architecture as currently framed at intake. The gap was real (verification gates were satisfiable but unsatisfied at intake). The structural intervention (staged commitment with verification gates) preserved Fund A’s capital optionality during verification without preventing full deployment when verification was achieved. The investment thesis itself proved sound; the architectural intervention was procedurally appropriate for the evidence position at decision time. (Claim Status: Inference for the hypothetical trajectory; the trajectory is demonstrative, not a prediction of any specific real engagement outcome.)
Section 8 — Document metadata (sample)
Report identifier: NPI-AR-SAMPLE-2026-001 (sample / anonymized)
Document version: 1.0
Initial publication date: May 23, 2026
Publisher: Novakian Paradigm Institute, Warsaw, Poland
Engagement type: Sample of Decision Architecture Review (institutional tier)
Methodology applied: Admissibility Check Protocol v1.0 per Novakian Paradigm Institute Methodology v1.0
Claim status standard applied: Novakian Paradigm Institute Claim Status Standard v1.0
Sample scope: Anonymized full deliverable. Fund A and Startup B are not specific real entities. Engagement structure, procedural application, and deliverable format are faithful to actual Admissibility Desk product.
Confidentiality: This sample is published openly. Actual Admissibility Reviews delivered under client engagement are covered by mutual confidentiality.
Estimated reading time: 30-40 minutes (full sample, English language).
PDF download: Available at canonical-URL/sample-admissibility-review.pdf upon completion of typeset layout.
Contact for actual engagement: contact@novakian.com with description of decision, claim, deployment, or commitment requiring review.
Sample ends.
What this sample demonstrates
This sample demonstrates the deliverable produced by an Admissibility Desk engagement at the standard institutional tier. The sample is approximately 6,500 words and follows the full structural format of an actual Review. The case is fully anonymized — neither Fund A nor Startup B refers to specific real entities, and the data presented does not reflect any specific real engagement.
What the sample shows: the engagement structure (intake, materials review, scope specification); the procedural application of the Admissibility Check Protocol v1.0 step by step; the per-claim analysis with verification gate specifications; the decision architecture analysis with scenario branches; the recommendations structure; the Evidence Ledger entries produced by the engagement; the document metadata format.
What the sample does not show: the conversation dynamics with the client (handled in actual engagement through structured intake calls and follow-up communication); the iterative refinement of scope (actual engagements typically include one to two rounds of scope clarification before Step 0 Silence Entry begins); the post-delivery follow-up calls (actual engagements include a structured one-hour debrief call within five business days of delivery and one optional follow-up call within thirty days).
When to request an actual engagement
Request an Admissibility Review when the requesting party faces a high-stakes pre-decision moment in the AI execution era and recognizes that the decision architecture itself warrants structural examination prior to commit. Typical engagement contexts include: investment funds preparing capital commitments at Series B and later stages; founders preparing major strategic pivots before public announcement; research teams preparing institutional publications with load-bearing claims; strategy groups examining AI deployment decisions before they become infrastructure; authors and analysts preparing claims that will enter the public record at scale.
The Institute does not perform Admissibility Reviews for: pre-revenue startups (insufficient claim base for institutional-threshold analysis); decisions whose entire scope is legal, financial, or technical due diligence (those work types are handled by specialist firms); decisions where the requesting party is seeking confirmation of an already-decided position (the Review is structural examination, not confirmation service); decisions where confidentiality cannot be mutually maintained.
Engagement tiers and pricing structure
The Institute currently offers three engagement tiers for Admissibility Desk work.
Single Signal Review. A bounded review of one specific claim, signal, or commitment. Two-to-three-week delivery timeline. Standard scope: one structured intake call, materials review, application of Admissibility Check Protocol v1.0, deliverable as written report. Pricing: $4,000-$8,000 USD depending on scope complexity. Suitable for analytical reviews of specific load-bearing claims in larger decision contexts where the rest of the decision architecture is being handled internally.
Decision Architecture Review. A full review of the decision architecture supporting an institutional commitment. Three-to-six-week delivery timeline. Standard scope (this sample represents this tier): structured intake with scope specification; materials review across multiple documents and data sources; full application of Admissibility Check Protocol v1.0; per-claim analysis with verification gates; decision architecture analysis with scenario branches; recommendations; structured one-hour debrief call and optional follow-up. Pricing: $18,000-$35,000 USD depending on scope complexity, document volume, and timeline urgency. Suitable for institutional commitments at significant scale where the Review materially affects commit/decline/restructure decisions.
Institutional Engagement. A standing arrangement for multiple Reviews over a defined period, typically twelve months. Custom scope per requesting party. Pricing: starting at $75,000 USD for twelve-month engagement with up to four Reviews of various scopes; pricing scales with engagement breadth. Suitable for institutional clients (funds, AI strategy teams, research groups) that anticipate recurring need for Review work and prefer standing arrangement over per-engagement contracting.
All tiers are covered by mutual confidentiality. The Institute does not publish client identities or engagement contents without explicit permission. The sample published on this page is the only public version of any Admissibility Review document.
Request process
To request an Admissibility Review, contact the Institute at contact@novakian.com with: brief description of the decision, claim, deployment, or commitment requiring review; intended tier (Single Signal, Decision Architecture, or Institutional Engagement); relevant timeline; non-confidential context that helps the Institute assess fit. The Institute reviews all serious inquiries within five business days and proposes engagement structure for those that meet engagement criteria. The Institute accepts engagements selectively; not every inquiry results in engagement.
