How does Featurize + Cluster work?

Featurize runs per-location feature-engineering: per-location size (square-footage + seat-count + transaction-volume + revenue + employees) + per-location vertical (NAICS code + per-vertical sub-segment) + per-location stage-of-lifecycle (new-store + growing + mature + declining + closing) + per-location geo (DMA + MSA + ZIP + latitude/longitude + urban/suburban/rural + Nielsen Scarborough + comScore + Resonate) + per-location demographic (per-location consumer-demographic via Acxiom + Experian + Equifax + Mosaic + PRIZM) + per-location competitive density (HHI + Lerner + per-3-mile + per-5-mile + per-10-mile competitor count) + per-location seasonality (per-location seasonal-index + per-location holiday-pattern) + per-location maturity (months-since-open + revenue-growth-trajectory) + per-location channel-mix (per-channel revenue-share + per-channel CAC). Per-feature standardization (z-score + min-max + robust scaler) + per-feature weight-vector. Cluster runs per-location peer-cohort assignment via per-method ensemble: k-NN nearest-neighbors (k tuned per-cohort) + hierarchical agglomerative clustering (Ward + single + complete + average linkage) + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + Mahalanobis distance. Per-cohort propensity-score matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + Inverse Probability of Treatment Weighting (IPTW) + nearest-neighbor matching + caliper matching for comparability.

What does Validate + Audit do?

Validate runs per-cohort stability + comparability + robustness verification. Per-cohort stability: silhouette score + Davies-Bouldin index + Calinski-Harabasz index + Adjusted Rand Index (ARI) + Normalized Mutual Information (NMI) + Fowlkes-Mallows + V-measure + Mutual Information + Homogeneity + Completeness + Variation of Information. Per-cohort comparability: per-cohort balance test + per-cohort standardized-mean-difference SMD (target SMD < 0.10) + per-cohort variance-ratio (target 0.5 < VR < 2.0) + per-cohort Wilcoxon rank-sum + per-cohort Kolmogorov-Smirnov + per-cohort Anderson-Darling + per-cohort propensity-score balance + per-cohort common-support overlap + per-cohort positivity assumption + per-cohort exchangeability test. Per-cohort robustness: per-cohort sensitivity analysis + per-cohort leave-one-out + per-cohort bootstrap + per-cohort permutation test + per-cohort cross-validation. Per-cohort severity classification: P0 SMD > 0.25 + propensity-score balance fail (cohort unusable) + P1 common-support overlap < 0.50 + 72-hour fix-window + P2 stability metric drift 7-day + P3 cohort-decomposition drift 30-day + P4 docs-only. Gate runs 5 anchors per-location per-cohort before any peer-cohort assignment distributes to benchmarking decision systems. (1) Per-location peer-cohort assignment algorithm + per-cohort stability + per-cohort comparability + per-cohort robustness + per-cohort feature-engineering. (2) FTC Section 5 + Pfizer 1972 + CFPB UDAAP + Lanham + USPTO + Robinson-Patman + FDD Item 19 financial performance representations when peer-cohort assignment shared with franchisees + 15-state franchise. (3) HIPAA + state mini-HIPAA + FINRA Rule 2210 + Rule 3110 + SEC Regulation FD + per-state professional licensing. (4) EU AI Act Article 50 transparency when AI-ML peer-cohort assignment + Article 13/14/15 + Annex III when AI-ML peer-cohort assignment drives benchmarking decisions + Article 6/27 FRIA + DSA + DMA + GDPR Article 6/7/22/28/30 + LGPD + DPDP + PIPEDA + Quebec Law 25 + CCPA + CPRA + 18-state. (5) WCAG 2.2 AA + ARIA + EAA + ADA Title III + Section 508 + SOX 302/404/906 + COSO + Exchange Act 13(b)(2) + FASB ASC 280 segment reporting + SEC Reg S-K. Audit writes a per-location per-cohort WORM assignment record: per-location feature snapshot + per-cohort assignment + per-cohort stability metric + per-cohort comparability + per-cohort robustness + per-anchor gate-pass + AI-ML provenance + EU AI Act FRIA. Retention: 7-year FTC + 7-year IRS + 7-year HIPAA + 7-year state bar + 6-year SEC + 3-year FINRA + 7-year SOX + GDPR Article 30 + EU AI Act Article 12 + SOC 2 CC7/CC8.

What does this skill connect to on the peer-cohort agent and across the swarm?

On the peer-cohort agent: per-location peer-cohort assignment + per-cohort stability + per-cohort comparability + per-cohort robustness. Across the swarm: per-location AI-calibrated forecasting (#600 DOWNSTREAM consumer of canonical peer-cohort) + per-location per-cohort two-sigma anomaly detection (#608 DOWNSTREAM consumer of canonical peer-cohort) + root-cause attribution sketch (#604 DOWNSTREAM consumer of per-cohort baseline) + per-location demographic feeds (UPSTREAM source of per-cohort demographic features) + integration-drift-monitor agent (#562 + #569 + #570) + per-state-overlay-composer (#599 UPSTREAM canonical for FDD Item 19 + per-state attorney advertising + FINRA per-state) + real-time change-event emission (#603 UPSTREAM canonical for per-cohort change-event) + per-field conflict-resolution policy (#607 same per-source authority hierarchy substrate). Commercial-pillar parent: /pre-emptive-churn-and-cohort-relative-trends.

What does the 6-workstream pre-engagement-baseline reporting cycle look like for this skill?

Every two weeks during the Tier 3 Fractional CMO with AI Swarm engagement, six workstreams report against the pre-engagement baseline. Workstream 1: per-portfolio per-location per-cohort peer-cohort-computation coverage — locations covered + cohorts computed + per-location features ingested. Workstream 2: Featurize per-location feature-engineering flow — per-location size + vertical + stage-of-lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix absorbed. Workstream 3: Cluster per-location peer-cohort assignment flow — k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + CEM + Genetic Matching + IPTW. Workstream 4: Validate per-cohort stability + comparability + robustness flow — silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI + SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + sensitivity analysis + bootstrap + permutation. Workstream 5: Regulatory-defense audit coverage — per-location peer-cohort + per-cohort stability + per-cohort comparability + per-cohort robustness + EU AI Act Article 50 + FDD Item 19 + FINRA + SOX + FASB ASC 280. Workstream 6: FBC feedback-loop pattern-learning — per-location per-cohort realized-vs-predicted assignment + per-cohort stability drift retrospective + per-cohort comparability retrospective.

Build pillar · peer-cohort agent

How to build peer-cohort computation at multi-location scale

Q: Why do scikit-learn + scipy.cluster + hdbscan + umap-learn + faiss + R MatchIt break at multi-location-multi-cohort-multi-period scale?

Each clustering + matching + dimensionality-reduction vendor ships per-account flat clustering primitive at single-method level. None coordinates per-location peer-cohort assignment against k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + Coarsened Exact Matching + Genetic Matching + IPTW simultaneously. None handles per-cohort stability metrics (silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI) + per-cohort comparability (SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + exchangeability) at the cross-cohort level. None gates against FDD Item 19 financial performance representations when peer-cohort assignment shared with franchisees + FINRA Rule 2210 when public-company peer-cohort + SOX 302/404/906 + FASB ASC 280 segment reporting. None writes a per-location per-cohort WORM assignment audit trail. The four-skill bundle Featurize + Cluster + Validate + Audit sits above the warehouse + clustering + matching + Bayesian-hierarchical substrate — it does not replace it.

Q: What does this skill connect to on the peer-cohort agent and across the swarm?

On the peer-cohort agent: per-location peer-cohort assignment + per-cohort stability + per-cohort comparability + per-cohort robustness. Across the swarm: per-location AI-calibrated forecasting (#600 DOWNSTREAM consumer of canonical peer-cohort) + per-location per-cohort two-sigma anomaly detection (#608 DOWNSTREAM consumer of canonical peer-cohort) + root-cause attribution sketch (#604 DOWNSTREAM consumer of per-cohort baseline) + per-location demographic feeds (UPSTREAM source of per-cohort demographic features) + integration-drift-monitor agent (#562 + #569 + #570) + per-state-overlay-composer (#599 UPSTREAM canonical for FDD Item 19 + per-state attorney advertising + FINRA per-state) + real-time change-event emission (#603 UPSTREAM canonical for per-cohort change-event) + per-field conflict-resolution policy (#607 same per-source authority hierarchy substrate). Commercial-pillar parent: /pre-emptive-churn-and-cohort-relative-trends.

Q: What does the 6-workstream pre-engagement-baseline reporting cycle look like for this skill?

Every two weeks during the Tier 3 Fractional CMO with AI Swarm engagement, six workstreams report against the pre-engagement baseline. Workstream 1: per-portfolio per-location per-cohort peer-cohort-computation coverage — locations covered + cohorts computed + per-location features ingested. Workstream 2: Featurize per-location feature-engineering flow — per-location size + vertical + stage-of-lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix absorbed. Workstream 3: Cluster per-location peer-cohort assignment flow — k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + CEM + Genetic Matching + IPTW. Workstream 4: Validate per-cohort stability + comparability + robustness flow — silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI + SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + sensitivity analysis + bootstrap + permutation. Workstream 5: Regulatory-defense audit coverage — per-location peer-cohort + per-cohort stability + per-cohort comparability + per-cohort robustness + EU AI Act Article 50 + FDD Item 19 + FINRA + SOX + FASB ASC 280. Workstream 6: FBC feedback-loop pattern-learning — per-location per-cohort realized-vs-predicted assignment + per-cohort stability drift retrospective + per-cohort comparability retrospective.

Snowflake + BigQuery + Databricks + dbt + Airflow + scikit-learn + scipy.cluster + hdbscan + umap-learn + opentsne + pynndescent + faiss + ScaNN + Annoy + HNSWLIB + scikit-learn AgglomerativeClustering (Ward + single + complete + average linkage) + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + ICA + LDA + Mahalanobis distance + DoWhy + EconML + causalml + Statsmodels + R MatchIt propensity-score-matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + Inverse Probability of Treatment Weighting (IPTW) + nearest-neighbor + caliper matching + PyMC + NumPyro + Stan + brms Bayesian hierarchical ship per- account flat clustering primitives. The Featurize + Cluster + Validate + Audit skill bundle on the peer- cohort agent sits above the warehouse + clustering + matching + Bayesian-hierarchical substrate and writes a per-location per-cohort canonical assignment record with named regulatory anchors covering per-location peer-cohort assignment + per-cohort stability (silhouette + Davies- Bouldin + Calinski-Harabasz + ARI + NMI) + per-cohort comparability (SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + exchangeability) + per-cohort robustness + EU AI Act Article 50 + FDD Item 19 + FINRA 2210 + SOX 302/404/906 + FASB ASC 280.

Published January 14, 2027 · 3,200 words

The 4-skill bundle on the peer-cohort agent

One agent. Four coordinated skills. The Featurize + Cluster + Validate + Audit bundle runs above the warehouse + clustering + matching + Bayesian-hierarchical substrate and writes one canonical per-location per-cohort assignment record.

Featurize

Per-location feature-engineering: per-location size + vertical (NAICS) + stage-of-lifecycle + geo (DMA + MSA + ZIP + lat/lon) + demographic (Acxiom + Experian + Mosaic + PRIZM) + competitive density (HHI + Lerner + per-3-mile + per-5-mile + per-10-mile competitor count) + seasonality + maturity + channel-mix. Per- feature standardization (z-score + min-max + robust scaler) + per-feature weight-vector.

Cluster

Per-location peer-cohort assignment via ensemble: k-NN + Ward + single + complete + average linkage + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + Mahalanobis distance. Per- cohort propensity-score matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + IPTW + nearest-neighbor + caliper matching.

Validate

Per-cohort stability: silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI + Fowlkes-Mallows + V- measure + Homogeneity + Completeness + Variation of Information. Per-cohort comparability: SMD (target < 0.10) + variance-ratio (0.5 < VR < 2.0) + Wilcoxon + Kolmogorov-Smirnov + Anderson-Darling + propensity-score balance + common-support overlap + positivity + exchangeability. Per-cohort robustness: sensitivity + leave-one-out + bootstrap + permutation + cross-validation. Per-cohort severity P0-P4.

Audit

Per-location per-cohort WORM assignment record: per- location feature snapshot + per-cohort assignment + per-cohort stability metric + per-cohort comparability + per-cohort robustness + per-anchor gate-pass + AI- ML provenance + EU AI Act FRIA. Retention: 7-year FTC + 7-year IRS + 7-year HIPAA + 7-year state bar + 6- year SEC + 3-year FINRA + 7-year SOX + GDPR Article 30 + EU AI Act Article 12 + SOC 2 CC7/CC8.

The real ecosystem this sits above

Featurize + Cluster + Validate + Audit does not replace warehouses, clustering libraries, matching packages, or Bayesian-hierarchical samplers. It sits above them and writes one canonical per-location per-cohort assignment record.

Clustering + nearest-neighbor

scikit-learn AgglomerativeClustering (Ward + linkage)
DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH
Gaussian Mixture Model + Bayesian GMM + DPMM
spectral clustering + Affinity Propagation
faiss + ScaNN + Annoy + HNSWLIB k-NN nearest-neighbor

Dimensionality + matching

UMAP + t-SNE + PCA + ICA + Kernel PCA + LDA
Mahalanobis distance + per-cohort feature-space
R MatchIt + DoWhy + EconML + causalml
Propensity-Score Matching + Coarsened Exact Matching
Genetic Matching + IPTW + caliper + nearest-neighbor

Bayesian hierarchical + warehouse

PyMC + NumPyro + Stan + brms Bayesian hierarchical
scikit-learn + XGBoost + LightGBM + CatBoost
Snowflake + BigQuery + Databricks + Redshift + ClickHouse
dbt + Dataform + SQLMesh + Airflow + Prefect + Dagster
Iceberg + Hudi + Delta Lake time-travel

Compliance overlay

Five anchors run per-location per-cohort before any peer- cohort assignment distributes to benchmarking decision systems. The first anchor is operationally distinctive: per-location peer-cohort assignment + per-cohort stability + per-cohort comparability + per-cohort robustness + per-cohort feature-engineering converge on every peer- cohort assignment decision.

Anchor 1: Per-location peer-cohort assignment + per- cohort stability + per-cohort comparability + per- cohort robustness + per-cohort feature-engineering (operationally distinctive)

Per-location peer-cohort assignment algorithm (k-NN nearest-neighbors + hierarchical agglomerative clustering Ward + single + complete + average linkage + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + ICA + Mahalanobis distance + per-cohort feature-space + per-cohort weight-vector + propensity-score matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + IPTW + nearest-neighbor matching + caliper matching). Per-cohort stability metric (silhouette + Davies-Bouldin + Calinski-Harabasz + Adjusted Rand Index (ARI) + Normalized Mutual Information (NMI) + Fowlkes-Mallows + V-measure + Mutual Information + Homogeneity + Completeness + Variation of Information). Per-cohort comparability metric (per-cohort balance test + per-cohort standardized-mean-difference SMD + per-cohort variance-ratio + per-cohort Wilcoxon rank-sum + per- cohort Kolmogorov-Smirnov + per-cohort Anderson- Darling + per-cohort propensity-score balance + per- cohort common-support overlap + per-cohort positivity assumption + per-cohort exchangeability test). Per- cohort robustness (per-cohort sensitivity analysis + per-cohort leave-one-out + per-cohort bootstrap + per-cohort permutation test + per-cohort cross- validation). Per-cohort feature-engineering (per- location size + vertical + stage-of-lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix).

Anchor 2: FTC + FDD Item 19 + Lanham

FTC Section 5 + Pfizer 1972 + CFPB UDAAP + Lanham + USPTO + Robinson-Patman + FDD Item 19 financial performance representations when peer-cohort assignment shared with franchisees + 15-state franchise.

Anchor 3: HIPAA + FINRA + per-vertical

HIPAA 45 CFR 164.502/504/514 + state mini-HIPAA + FINRA Rule 2210 + Rule 3110 + SEC Regulation FD + per-state professional licensing.

Anchor 4: EU AI Act + AI-ML peer-cohort assignment

EU AI Act Article 50 transparency when AI-ML peer- cohort assignment + Article 13/14/15 + Annex III when AI-ML peer-cohort assignment drives benchmarking decisions + Article 6/27 FRIA + DSA + DMA. GDPR Article 6/7/22/28/30 + LGPD + DPDP + PIPEDA + Quebec Law 25 + CCPA + CPRA + 18-state.

Anchor 5: Accessibility + SOX + FASB + WORM retention

WCAG 2.2 AA + ARIA + EAA + ADA Title III + Section 508. SOX 302/404/906 + COSO + Exchange Act 13(b)(2) + FASB ASC 280 segment reporting + SEC Reg S-K. NIST AI RMF + ISO 42001 + ISO 27001 + SOC 2 Type II. Per-vendor LLM zero-retention + per-source DPA + per-API rate- limit. Storage: AWS S3 Object Lock + Azure Blob immutable + GCS + Wasabi WORM. Retention: 7-year FTC + 7-year IRS + 7-year HIPAA + 7-year state bar + 6- year SEC + 3-year FINRA + 7-year SOX + GDPR Article 30 + EU AI Act Article 12 + SOC 2 CC7/CC8.

6-workstream reporting cycle

Every two weeks during a Tier 3 Fractional CMO engagement, six workstreams report against the pre-engagement baseline. No peer-cohort accuracy claims. Process commitments only.

1. Per-portfolio per-location per-cohort peer- cohort-computation coverage. Locations covered + cohorts computed + per-location features ingested.
2. Featurize per-location feature-engineering flow. Per-location size + vertical + stage-of- lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix absorbed.
3. Cluster per-location peer-cohort assignment flow. k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + CEM + Genetic Matching + IPTW.
4. Validate per-cohort stability + comparability + robustness flow. Silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI + SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + sensitivity + bootstrap + permutation.
5. Regulatory-defense audit coverage. Per-location peer-cohort + per-cohort stability + per- cohort comparability + per-cohort robustness + EU AI Act Article 50 + FDD Item 19 + FINRA + SOX + FASB ASC 280.
6. FBC feedback-loop pattern-learning. Per-location per-cohort realized-vs-predicted assignment + per-cohort stability drift retrospective + per-cohort comparability retrospective.

FAQ

What is peer-cohort computation at multi-location scale — and what is the per-location-peer-cohort-assignment-times-per-cohort-stability-times-per-cohort-comparability-times-propensity-score-matching-times-replication-crisis-discipline problem distinctive to this skill?: A multi-location operator with 50-300 stores ships per-location peer-cohort assignment for benchmarking + forecasting + anomaly-detection + root-cause-attribution + market-expansion-decision. Naive "all stores" benchmarking masks per-location variance (urban vs suburban + mature vs new + high-density-competition vs low-density + high-AOV-vertical vs low-AOV + seasonal vs non-seasonal). The four-skill bundle on the peer-cohort agent — Featurize, Cluster, Validate, Audit — sits above the warehouse + clustering + matching + Bayesian-hierarchical substrate (Snowflake + BigQuery + Databricks + dbt + Airflow + scikit-learn + scipy.cluster + hdbscan + umap-learn + faiss + ScaNN + scikit-learn AgglomerativeClustering + DBSCAN + HDBSCAN + Gaussian Mixture Model + spectral clustering + UMAP + t-SNE + PCA + Mahalanobis + DoWhy + EconML + causalml + R MatchIt propensity-score-matching + Coarsened Exact Matching + Genetic Matching + IPTW + PyMC + NumPyro + Stan + brms) and writes a per-location per-cohort canonical assignment record. The operationally distinctive anchor: per-location peer-cohort assignment algorithm (k-NN nearest-neighbors + hierarchical agglomerative clustering Ward + single + complete + average linkage + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + ICA + Mahalanobis distance + per-cohort feature-space + per-cohort weight-vector + propensity-score matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + Inverse Probability of Treatment Weighting (IPTW) + nearest-neighbor matching + caliper matching) + per-cohort stability metric (silhouette score + Davies-Bouldin index + Calinski-Harabasz index + Adjusted Rand Index (ARI) + Normalized Mutual Information (NMI) + Fowlkes-Mallows + V-measure + Mutual Information + Homogeneity + Completeness + Variation of Information) + per-cohort comparability metric (per-cohort balance test + per-cohort standardized-mean-difference SMD + per-cohort variance-ratio + per-cohort Wilcoxon rank-sum + per-cohort Kolmogorov-Smirnov + per-cohort Anderson-Darling + per-cohort propensity-score balance + per-cohort common-support overlap + per-cohort positivity assumption + per-cohort exchangeability test) + per-cohort robustness (sensitivity analysis + leave-one-out + bootstrap + permutation test + cross-validation) + per-cohort feature-engineering (per-location size + vertical + stage-of-lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix).
Why do scikit-learn + scipy.cluster + hdbscan + umap-learn + faiss + R MatchIt break at multi-location-multi-cohort-multi-period scale?: Each clustering + matching + dimensionality-reduction vendor ships per-account flat clustering primitive at single-method level. None coordinates per-location peer-cohort assignment against k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + Coarsened Exact Matching + Genetic Matching + IPTW simultaneously. None handles per-cohort stability metrics (silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI) + per-cohort comparability (SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + exchangeability) at the cross-cohort level. None gates against FDD Item 19 financial performance representations when peer-cohort assignment shared with franchisees + FINRA Rule 2210 when public-company peer-cohort + SOX 302/404/906 + FASB ASC 280 segment reporting. None writes a per-location per-cohort WORM assignment audit trail. The four-skill bundle Featurize + Cluster + Validate + Audit sits above the warehouse + clustering + matching + Bayesian-hierarchical substrate — it does not replace it.
How does Featurize + Cluster work?: Featurize runs per-location feature-engineering: per-location size (square-footage + seat-count + transaction-volume + revenue + employees) + per-location vertical (NAICS code + per-vertical sub-segment) + per-location stage-of-lifecycle (new-store + growing + mature + declining + closing) + per-location geo (DMA + MSA + ZIP + latitude/longitude + urban/suburban/rural + Nielsen Scarborough + comScore + Resonate) + per-location demographic (per-location consumer-demographic via Acxiom + Experian + Equifax + Mosaic + PRIZM) + per-location competitive density (HHI + Lerner + per-3-mile + per-5-mile + per-10-mile competitor count) + per-location seasonality (per-location seasonal-index + per-location holiday-pattern) + per-location maturity (months-since-open + revenue-growth-trajectory) + per-location channel-mix (per-channel revenue-share + per-channel CAC). Per-feature standardization (z-score + min-max + robust scaler) + per-feature weight-vector. Cluster runs per-location peer-cohort assignment via per-method ensemble: k-NN nearest-neighbors (k tuned per-cohort) + hierarchical agglomerative clustering (Ward + single + complete + average linkage) + DBSCAN + HDBSCAN + OPTICS + Mean Shift + BIRCH + Gaussian Mixture Model + Bayesian GMM + Dirichlet Process Mixture + spectral clustering + Affinity Propagation + UMAP + t-SNE + PCA + Mahalanobis distance. Per-cohort propensity-score matching (PSM) + Coarsened Exact Matching (CEM) + Genetic Matching (GenMatch) + Inverse Probability of Treatment Weighting (IPTW) + nearest-neighbor matching + caliper matching for comparability.
What does Validate + Audit do?: Validate runs per-cohort stability + comparability + robustness verification. Per-cohort stability: silhouette score + Davies-Bouldin index + Calinski-Harabasz index + Adjusted Rand Index (ARI) + Normalized Mutual Information (NMI) + Fowlkes-Mallows + V-measure + Mutual Information + Homogeneity + Completeness + Variation of Information. Per-cohort comparability: per-cohort balance test + per-cohort standardized-mean-difference SMD (target SMD < 0.10) + per-cohort variance-ratio (target 0.5 < VR < 2.0) + per-cohort Wilcoxon rank-sum + per-cohort Kolmogorov-Smirnov + per-cohort Anderson-Darling + per-cohort propensity-score balance + per-cohort common-support overlap + per-cohort positivity assumption + per-cohort exchangeability test. Per-cohort robustness: per-cohort sensitivity analysis + per-cohort leave-one-out + per-cohort bootstrap + per-cohort permutation test + per-cohort cross-validation. Per-cohort severity classification: P0 SMD > 0.25 + propensity-score balance fail (cohort unusable) + P1 common-support overlap < 0.50 + 72-hour fix-window + P2 stability metric drift 7-day + P3 cohort-decomposition drift 30-day + P4 docs-only. Gate runs 5 anchors per-location per-cohort before any peer-cohort assignment distributes to benchmarking decision systems. (1) Per-location peer-cohort assignment algorithm + per-cohort stability + per-cohort comparability + per-cohort robustness + per-cohort feature-engineering. (2) FTC Section 5 + Pfizer 1972 + CFPB UDAAP + Lanham + USPTO + Robinson-Patman + FDD Item 19 financial performance representations when peer-cohort assignment shared with franchisees + 15-state franchise. (3) HIPAA + state mini-HIPAA + FINRA Rule 2210 + Rule 3110 + SEC Regulation FD + per-state professional licensing. (4) EU AI Act Article 50 transparency when AI-ML peer-cohort assignment + Article 13/14/15 + Annex III when AI-ML peer-cohort assignment drives benchmarking decisions + Article 6/27 FRIA + DSA + DMA + GDPR Article 6/7/22/28/30 + LGPD + DPDP + PIPEDA + Quebec Law 25 + CCPA + CPRA + 18-state. (5) WCAG 2.2 AA + ARIA + EAA + ADA Title III + Section 508 + SOX 302/404/906 + COSO + Exchange Act 13(b)(2) + FASB ASC 280 segment reporting + SEC Reg S-K. Audit writes a per-location per-cohort WORM assignment record: per-location feature snapshot + per-cohort assignment + per-cohort stability metric + per-cohort comparability + per-cohort robustness + per-anchor gate-pass + AI-ML provenance + EU AI Act FRIA. Retention: 7-year FTC + 7-year IRS + 7-year HIPAA + 7-year state bar + 6-year SEC + 3-year FINRA + 7-year SOX + GDPR Article 30 + EU AI Act Article 12 + SOC 2 CC7/CC8.
What does this skill connect to on the peer-cohort agent and across the swarm?: On the peer-cohort agent: per-location peer-cohort assignment + per-cohort stability + per-cohort comparability + per-cohort robustness. Across the swarm: per-location AI-calibrated forecasting (#600 DOWNSTREAM consumer of canonical peer-cohort) + per-location per-cohort two-sigma anomaly detection (#608 DOWNSTREAM consumer of canonical peer-cohort) + root-cause attribution sketch (#604 DOWNSTREAM consumer of per-cohort baseline) + per-location demographic feeds (UPSTREAM source of per-cohort demographic features) + integration-drift-monitor agent (#562 + #569 + #570) + per-state-overlay-composer (#599 UPSTREAM canonical for FDD Item 19 + per-state attorney advertising + FINRA per-state) + real-time change-event emission (#603 UPSTREAM canonical for per-cohort change-event) + per-field conflict-resolution policy (#607 same per-source authority hierarchy substrate). Commercial-pillar parent: /pre-emptive-churn-and-cohort-relative-trends.
What does the 6-workstream pre-engagement-baseline reporting cycle look like for this skill?: Every two weeks during the Tier 3 Fractional CMO with AI Swarm engagement, six workstreams report against the pre-engagement baseline. Workstream 1: per-portfolio per-location per-cohort peer-cohort-computation coverage — locations covered + cohorts computed + per-location features ingested. Workstream 2: Featurize per-location feature-engineering flow — per-location size + vertical + stage-of-lifecycle + geo + demographic + competitive density + seasonality + maturity + channel-mix absorbed. Workstream 3: Cluster per-location peer-cohort assignment flow — k-NN + Ward + DBSCAN + HDBSCAN + GMM + spectral + UMAP + Mahalanobis + propensity-score matching + CEM + Genetic Matching + IPTW. Workstream 4: Validate per-cohort stability + comparability + robustness flow — silhouette + Davies-Bouldin + Calinski-Harabasz + ARI + NMI + SMD + variance-ratio + Kolmogorov-Smirnov + propensity-score balance + common-support + positivity + sensitivity analysis + bootstrap + permutation. Workstream 5: Regulatory-defense audit coverage — per-location peer-cohort + per-cohort stability + per-cohort comparability + per-cohort robustness + EU AI Act Article 50 + FDD Item 19 + FINRA + SOX + FASB ASC 280. Workstream 6: FBC feedback-loop pattern-learning — per-location per-cohort realized-vs-predicted assignment + per-cohort stability drift retrospective + per-cohort comparability retrospective.

Engage Completions

Two ways to engage. The Tier 1 AI Readiness Assessment maps the warehouse + clustering + matching + Bayesian- hierarchical substrate + per-location peer-cohort + per- cohort stability + per-cohort comparability + per-cohort robustness surface against the Featurize + Cluster + Validate + Audit bundle. The Tier 3 Fractional CMO with AI Swarm embeds 1-2 days per week for 6+ months and runs the bundle end-to-end against the peer-cohort agent across the swarm.

Tier 1 AI Readiness Assessment — $10k, 2-3 weeks Tier 3 Fractional CMO with AI Swarm — $15-25k/mo Not sure which tier? Take the 3-question quiz