§ Evidence

Evaluated across 1,035 regulated documents.

At DIA 2026, Clinplex presented a multi-framework gap-detection methodology evaluated across SOPs, CAPAs, deviations, audit records, CMC submissions, validation packages and preclinical documents. 96% gap-detection accuracy against blinded SME review.

DIA 2026 · Abstract 115882 · Track 8: Regulatory AI · Poster P312
View the DIA abstract → Request the presented poster
96%

Gap-detection accuracy

Against SME-panel ground truth.

3.0×

More gaps surfaced

12.3 gaps per SOP vs 4.1 under single-framework review.

73%

Less assessment time

Cross-framework regulatory assessment.

98%

Critical-tier sensitivity

Agreement with SME review on Critical findings.

§ 1 · Method

Human review was part of the benchmark.

A 1,035-document regulated corpus from biologics, small-molecule and CDMO operating environments. A stratified 312-document subset was independently reviewed by a seven-person Quality and Regulatory panel, blinded to Clinplex output.

1,035 documents→Clinplex evaluation→Blinded SME review→Consensus adjudication
7
SME reviewers
Quality and Regulatory subject-matter experts.
312
Stratified review subset
Documents independently reviewed by the panel.
κ = 0.87
Inter-rater agreement
Cohen's κ for AI-vs-SME agreement, with 95% CI.
SOPsCAPAsDeviationsAudit recordsCMC submissionsIQ / OQ / PQ validation packagesPreclinical study reports
§ 2 · Results

Multi-framework analysis found substantially more.

PerformanceSingle-framework reviewClinplex multi-framework
Average gaps per SOP4.112.3
Critical-tier sensitivity0.710.98
Specificity0.880.94
Cross-document linkageNot evaluatedSOP → CAPA mapped
13 additional cross-regulatory gaps were identified that single-framework assessment did not capture.
§ 3 · Why the Difference

Why 3× more findings?

A conventional review asks
"Does this SOP meet 21 CFR 211?"
Clinplex asks
"Does this SOP satisfy the combined obligations of 21 CFR 211, ICH Q7/Q9/Q10, EU GMP and WHO, and the related records that implement them?"
Evaluated: FDA 21 CFR 210/211 · ICH Q7 / Q9 / Q10 · EU GMP Annexes · WHO TRS · 21 CFR Part 11 · ICH Q9-aligned risk classification
§ 4 · Rigor

Consistency, and its limits.

Reproducibility

  • Outputs evaluated for stability across repeated runs.
  • Controls for operating-environment variance.
  • Test / retest consistency evaluated over time, not from a single run.

Limitations

The evaluation covered selected organizations, regulatory classes and document types. Results should not be read as universal performance across every organization, document type or framework. Broader cross-organizational and longitudinal evaluation is ongoing.

Reproducibility protocol and evaluation artifacts are available under NDA.

§ 5 · What a Finding Contains

The full evidentiary chain.

01Source evidenceThe record and context supporting the finding
02Regulatory basisThe governing requirement, cited
03Impact and rationaleWhy it matters and what related evidence is affected
04Proposed remediationA review-ready path to address the finding
Request a redacted sample analysis →
§ 6 · The Record

Review the underlying work.

DIA 2026 Global Annual Meeting · Abstract 115882 · Poster P312

AI-Enabled Multi-Framework Regulatory and Quality Gap Detection for Pharmaceutical Operations Across the Product Lifecycle

Claims register: every published figure, its definition, denominator and source CPX-CLAIMS-001 · v1.1
Evaluation claims
WordingDefinitionDenominatorSourceVerifiedStatus
"96% gap-detection accuracy"Proportion of AI-identified gaps confirmed as true gaps by the blinded SME panel. Reported with 95% CI.312-document subsetDIA 2026 poster P3122026-06-09Verified
"12.3 gaps per SOP vs 4.1" / "3.0× more gaps surfaced"Mean gaps per SOP under multi-framework evaluation vs single-framework manual review of the same SOPs. 3.0× is the ratio 12.3 / 4.1.SOP subsetDIA 2026 poster P3122026-06-09Verified
"73% less cross-framework assessment time"Reduction in time to complete cross-framework assessment vs manual single-framework baseline.Timed subsetDIA 2026 poster P3122026-06-09Verified
"κ = 0.87 inter-rater agreement"Cohen's κ for agreement between AI findings and SME-panel adjudication. Reported with 95% CI.312-document subsetDIA 2026 poster P3122026-06-09Verified
"98% critical-tier sensitivity"Proportion of SME-adjudicated Critical findings detected by Clinplex.Critical-tier, 312-document subsetDIA 2026 poster P3122026-06-09Verified
"Evaluated on a 1,035-document corpus"Total documents in the evaluation corpus. "Evaluated," not "validated": a benchmark evaluation, not a validation under a defined protocol with acceptance criteria.n = 1,035DIA 2026 poster P3122026-06-09Verified
"Blinded 312-document subset" / "7 SME reviewers"Stratified subset independently reviewed by a 7-person Quality and Regulatory panel, blinded to AI output; disputes resolved by consensus.n = 312 of 1,035DIA 2026 poster P3122026-06-09Verified
"273 total gaps"Unique gaps identified across the full corpus. 12.3 per SOP is the SOP-subset mean; 273 is the corpus-wide count. Different denominators, same study.Full corpusDIA 2026 poster P3122026-06-09Verified
Product and business claims
WordingDefinitionDenominatorSourceVerifiedStatus
"Findings come back within 48 hours"Inspection Stress Test turnaround from receipt of the three records to delivery of findings, in business hours.Per engagementOperational commitment2026-08-31Verified
"Mutual NDA goes out the same day"NDA dispatched within one business day of request.Per requestOperational commitment2026-08-31Verified
"Four lifecycle phases"Discovery & preclinical, quality/manufacturing, clinical operations, CMC/regulatory submissions, as shown in the product.ProductProduct2026-08-31Verified
"74% of 202 FDA CRLs involved quality or manufacturing deficiencies"Share of the 202 complete response letters published by FDA citing quality or manufacturing deficiencies.n = 202 published CRLsopenFDA + trade analysis2026-08-31Verified

Only current, verified claims are listed. Owner: security@clinplex.com. Superseded claims are removed from this register and from the site.

Now run it on
your own records.

One closed deviation, its CAPA and the governing SOP. Ranked gaps with citations and a drafted remediation. Free.

Run an Inspection Stress Test → For regulators →