Done-for-you offer · Fractional CMO with AI Swarm · internal-link-recommendation-engine skill
Completions builds the per-page internal-link recommendation engine — closed loop with loop 16 link-equity diagnostic
You operate 10,000-1,000,000 pages across 50-1,500 locations with internal-link equity that decays from orphan pages, redirect chains, deep nesting, and weak anchor distribution. Your SEO team demands per-page per-source per-target recommendations. Your CMO demands per-recommendation impact -projection proof. Your counsel demands FTC Section 5 + FTC Endorsement Guides + ADA Title III + state-AG + per-vertical regulated-content compliance on every link. Completions builds the internal-link-recommendation-engine skill on the internal-link-orchestration agent end-to-end with per-page per-source per-target link recommendation + per-recommendation per-anchor-text generation + per-recommendation per-equity -impact projection + per-recommendation feedback to link -equity diagnostic. You keep every artifact. You keep the page graph + recommendation log + anchor-text registry + impact -projection model. You keep the ability to in-house at any time.
Published September 24, 2026
What we operate every week
Per-page per-source per-target internal-link recommendation across 10,000-1,000,000 pages × per-page per-source per -target candidate-pair scoring via 15+ embedding models (BERT + Sentence-T5 + E5 + cohere-embed + voyage-embed + jina-embeddings + bge-large + arctic-embed + nomic-embed + gemini-embedding + UAE-large + multilingual-e5 + GIST -embedding + mxbai-embed + gte-large) + per-pair 7-dimension scoring (topical-relevance + user-journey-progression + conversion-path + link-equity-flow + PageRank-impact + cannibalization-risk).
Per-recommendation per-anchor-text generation via per-target keyword universe (primary + secondary + supporting + entity + brand + topical-cluster) with per-anchor-text variation (exact-match + phrase-match + partial-match + branded + naked-URL + generic + image-alt + co-occurrence) + per -anchor-text diversity enforcement + Penguin-safety enforcement + Helpful-Content-Update-safety enforcement + user-intent alignment + brand-voice alignment.
Per-recommendation per-equity-impact projection emits per -recommendation projected-PageRank-delta + projected-rank -delta + projected-traffic-delta + projected-conversion -delta + projected-revenue-delta + projected-time-to-impact + projected-confidence-tier + projected-explainability.
Per-recommendation feedback to link-equity diagnostic emits per-recommendation post-implementation actual-deltas into the link-equity diagnostic model with per-recommendation Bayesian-updating + counterfactual-validation + propensity -score-matching + experimental-design treatment-assignment -protocol.
Closed feedback loop coordination across loop 42 + loop 16 + url-hierarchy-authoring + on-demand-cannibalization-risk -scoring + per-location-link-sculpting. Per-vertical compliance overlay (FTC Section 5 + FTC Endorsement Guides + ADA Title III + state-AG + per-vertical regulated -content).
Why in-house breaks at multi-location site scale
Per-page per-source per-target recommendation across 10,000 -1,000,000 pages × 15+ embedding models × 7 scoring dimensions requires production ML infrastructure. Per -recommendation per-anchor-text generation across 8 variation types requires production generation infrastructure. Per-recommendation per-equity-impact projection requires production projection infrastructure with causal-uplift methodology. Per-recommendation feedback to link-equity diagnostic requires data-science capacity with closed-feedback-loop expertise. Closed-feedback-loop coordination across loop 42 + loop 16 requires orchestration capacity. Per-vertical compliance overlay requires legal-engineering capacity. Per-page graph maintenance requires production maintenance capacity.
Completions absorbs all seven axes under one Tier 3 Fractional CMO with AI Swarm engagement.
How the engagement progresses
Tier 1 AI Readiness Assessment ($10k, 2-3 weeks, diagnostic). Audits the current operation across seven axes. Deliverable: gap-pack report with per-page link -equity exposure estimate + per-recommendation impact -projection accuracy estimate.
Tier 2 AI Swarm Setup Sprint ($25-50k, 4-8 weeks, build with 30-day operating tail). Builds the per -page internal-link recommendation engine on operator infrastructure — internal-link-recommendation-engine + per -page-per-source-per-target-recommendation + per -recommendation-anchor-text-generation + per-recommendation -equity-impact-projection + per-recommendation-feedback-to -diagnostic on internal-link-orchestration agent + link -equity-diagnostic + url-hierarchy-authoring on url -hierarchy-authoring + on-demand-cannibalization-risk -scoring on cross-location-cannibalization-defense + per -location-link-sculpting on link-sculpting-at-scale.
Tier 3 Fractional CMO with AI Swarm ($15-25k/ month, 6-month minimum, 1-2 days/wk embedded). Continues operating the engine with weekly per-page graph refresh + monthly per-recommendation generation refresh + quarterly per-recommendation impact-projection refresh + per-event per-recommendation feedback to link-equity diagnostic + cross-agent swarm coordination.
Frequently asked
What does "Completions builds the per-page internal-link recommendation engine — closed loop with loop 16 link-equity diagnostic" actually deliver?
Completions builds and operates per-page per-source per-target internal-link recommendation + per-recommendation per-anchor-text generation + per-recommendation per-equity-impact projection + per-recommendation feedback to link-equity diagnostic across the operator multi-location site surface. Per-page per-source per-target internal-link recommendation across 10,000-1,000,000 pages × per-page per-source per-target candidate-pair scoring via semantic similarity (BERT + Sentence-T5 + E5 + cohere-embed + voyage-embed + jina-embeddings + bge-large + arctic-embed + nomic-embed + gemini-embedding + UAE-large + multilingual-e5 + GIST-embedding + mxbai-embed + gte-large) + per-pair topical-relevance scoring + per-pair user-journey-progression scoring + per-pair conversion-path scoring + per-pair link-equity-flow projection + per-pair PageRank-impact projection + per-pair cannibalization-risk scoring. Per-recommendation per-anchor-text generation via per-target keyword universe (per-target primary-keyword + per-target secondary-keyword + per-target supporting-keyword + per-target entity + per-target brand + per-target topical-cluster) with per-anchor-text variation (exact-match + phrase-match + partial-match + branded + naked-URL + generic + image-alt + co-occurrence) + per-anchor-text diversity enforcement + per-anchor-text Penguin-safety enforcement + per-anchor-text Helpful-Content-Update-safety enforcement + per-anchor-text user-intent alignment + per-anchor-text brand-voice alignment. Per-recommendation per-equity-impact projection emits per-recommendation projected-PageRank-delta + per-recommendation projected-rank-delta + per-recommendation projected-traffic-delta + per-recommendation projected-conversion-delta + per-recommendation projected-revenue-delta + per-recommendation projected-time-to-impact + per-recommendation projected-confidence-tier + per-recommendation projected-explainability. Per-recommendation feedback to link-equity diagnostic emits per-recommendation per-source per-target post-implementation actual-PageRank-delta + actual-rank-delta + actual-traffic-delta + actual-conversion-delta + actual-revenue-delta + actual-time-to-impact + actual-confidence-tier into the link-equity diagnostic model with per-recommendation Bayesian-updating + per-recommendation counterfactual-validation + per-recommendation propensity-score-matching + per-recommendation experimental-design treatment-assignment-protocol. Closed feedback loop coordination across loop 42 (internal-link-recommendation-engine) + loop 16 (link-equity-diagnostic) + url-hierarchy-authoring + on-demand-cannibalization-risk-scoring + per-location-link-sculpting. Per-vertical compliance overlay (FTC Section 5 disclosure-link requirement + FTC Endorsement Guides affiliate-link disclosure + ADA Title III accessible-link requirement + state-attorney-general advertising-claim link-target + per-vertical regulated-content link-target). Operator team owns the page graph + recommendation log + anchor-text registry + impact-projection model. Completions owns the swarm orchestration on the internal-link-orchestration agent.
Why does in-house per-page internal-link recommendation break at multi-location site scale?
In-house operation at multi-location site scale fails on seven axes: (1) per-page per-source per-target internal-link recommendation across 10,000-1,000,000 pages × per-page candidate-pair scoring via 15+ embedding models × per-pair 7 scoring dimensions requires production ML infrastructure with model-routing + retry + idempotency unstaffable by internal teams; (2) per-recommendation per-anchor-text generation across 8 anchor-text variation types × per-anchor-text diversity enforcement × per-anchor-text Penguin-safety enforcement × per-anchor-text Helpful-Content-Update-safety enforcement requires production generation infrastructure with per-anchor-text governance; (3) per-recommendation per-equity-impact projection across 7 projected-delta dimensions × per-recommendation confidence-tier × per-recommendation explainability requires production projection infrastructure with causal-uplift methodology; (4) per-recommendation feedback to link-equity diagnostic with per-recommendation Bayesian-updating + counterfactual-validation + propensity-score-matching + experimental-design treatment-assignment-protocol requires data-science capacity with closed-feedback-loop expertise; (5) closed-feedback-loop coordination across loop 42 + loop 16 + url-hierarchy-authoring + on-demand-cannibalization-risk-scoring + per-location-link-sculpting requires orchestration capacity; (6) per-vertical compliance overlay covering FTC Section 5 + FTC Endorsement Guides + ADA Title III + state-AG + per-vertical regulated-content requires legal-engineering capacity; (7) per-page graph maintenance (per-page URL change + per-page deletion + per-page redirect + per-page noindex + per-page canonical change × per-week refresh) requires production maintenance capacity. Completions absorbs all seven axes under one Tier 3 Fractional CMO with AI Swarm engagement.
What does the engagement look like across Tier 1 → Tier 2 → Tier 3?
Tier 1 AI Readiness Assessment ($10k, 2-3 weeks, diagnostic): Completions audits the operator current internal-link recommendation operation across seven axes — per-page per-source per-target recommendation coverage + per-recommendation per-anchor-text generation maturity + per-recommendation per-equity-impact projection + per-recommendation feedback-to-diagnostic + closed-feedback-loop coordination + per-vertical compliance overlay + per-page graph maintenance. Deliverable: gap-pack report with per-page link-equity exposure estimate + per-recommendation impact-projection accuracy estimate. Tier 2 AI Swarm Setup Sprint ($25-50k, 4-8 weeks, build with 30-day operating tail): Completions builds the per-page internal-link recommendation engine on operator infrastructure — internal-link-recommendation-engine + per-page-per-source-per-target-recommendation + per-recommendation-anchor-text-generation + per-recommendation-equity-impact-projection + per-recommendation-feedback-to-diagnostic on internal-link-orchestration agent + link-equity-diagnostic on internal-link-orchestration + url-hierarchy-authoring on url-hierarchy-authoring + on-demand-cannibalization-risk-scoring on cross-location-cannibalization-defense + per-location-link-sculpting on link-sculpting-at-scale. Tier 3 Fractional CMO with AI Swarm ($15-25k/month, 6-month minimum, 1-2 days/wk embedded): Completions continues operating the engine with weekly per-page graph refresh + monthly per-recommendation generation refresh + quarterly per-recommendation impact-projection refresh + per-event per-recommendation feedback to link-equity diagnostic + cross-agent swarm coordination.
Who owns the page graph, recommendation log, anchor-text registry, and impact-projection model?
Operator owns 100% of every artifact: page graph (in operator data infrastructure — Snowflake + Databricks + BigQuery + Redshift + Postgres operator data warehouse + operator graph database — Neo4j + ArangoDB + TigerGraph + JanusGraph + Amazon Neptune + Memgraph + Dgraph + RedisGraph), per-recommendation log (in operator repo with WORM-storage for audit trail), per-anchor-text registry (in operator repo with per-target keyword universe + per-anchor-text variation taxonomy + per-anchor-text Penguin-safety enforcement rules + per-anchor-text Helpful-Content-Update-safety enforcement rules), per-recommendation impact-projection model code (in operator repo with operator-controlled deploy pipeline + operator-data-science-team-aligned), per-recommendation feedback-to-diagnostic model code (operator-owned + operator-data-science-team-aligned), per-vertical compliance overlay (rule library in operator repo with attorney-approved updates), FTC Section 5 + FTC Endorsement Guides + ADA Title III + state-AG + per-vertical regulated-content disclosure register (operator-owned + operator-counsel-maintained), per-page graph crawler credentials (Screaming Frog + Sitebulb + DeepCrawl + Botify + ContentKing + OnCrawl + Lumar + Ahrefs Site Audit + SEMrush Site Audit + Moz Site Audit under operator billing + operator credentials), brand spec (versioned in operator repo), LLM prompts (in operator repo), audit trail (retention infrastructure on operator cloud account). Completions owns: the orchestration knowledge — how to design per-page per-source per-target recommendation contracts + how to tune per-recommendation per-anchor-text generation + how to debug per-recommendation impact-projection cascades + how to coordinate the closed loop with link-equity-diagnostic + url-hierarchy-authoring + on-demand-cannibalization-risk-scoring + per-location-link-sculpting siblings. The operator can in-house at any time; Completions credentials revoke immediately on engagement-end.
What KPIs will Completions commit to on Tier 3 engagement?
Typical Tier 3 commitments: (1) per-page per-source per-target recommendation coverage at 99-percent target across the operator page graph; (2) per-recommendation per-anchor-text generation accuracy at 95-percent target measured against SEO-team-validated golden set; (3) per-recommendation per-equity-impact projection accuracy at 80-percent target measured against post-implementation actual-delta; (4) per-recommendation feedback-to-diagnostic integration latency under 24-hour end-to-end; (5) per-page link-equity-flow improvement at 20-50-percent target (operator-baseline-dependent); (6) per-page organic-traffic-lift at 10-30-percent target (operator-baseline-dependent); (7) per-page anchor-text Penguin-safety adherence at 100-percent target; (8) per-page anchor-text Helpful-Content-Update-safety adherence at 100-percent target; (9) per-vertical compliance overlay coverage at 99.9-percent target; (10) closed-feedback-loop coordination latency under 2-second end-to-end. Each KPI measured against pre-engagement baseline.
How does engagement end and what is the operator transition path?
Tier 3 engagements are 6-month minimum with 90-day notice. At engagement end, Completions transitions the per-page internal-link recommendation engine operation back to operator in-house in 30-60 days: operating-playbook hand-off + in-house staff training across 3-5 operator team members covering per-page per-source per-target recommendation + per-recommendation per-anchor-text generation + per-recommendation per-equity-impact projection + per-recommendation feedback-to-diagnostic + closed-feedback-loop coordination + per-vertical compliance overlay management + per-page graph maintenance + cross-agent coordination + page graph infrastructure hand-off + per-recommendation log hand-off + per-anchor-text registry hand-off + per-recommendation impact-projection model code hand-off + per-recommendation feedback-to-diagnostic model code hand-off + per-page graph crawler credentials hand-off + LLM prompts hand-off + audit trail hand-off; Completions credentials revoke immediately on engagement-end. Operator can re-engage Completions at any time on Tier 1 or Tier 2 cadence.
Engage Completions
Start with the AI Readiness Assessment (Tier 1, 2-3 weeks, $10k). Hand off to Tier 2 ($25-50k, 4-8 weeks) for the build. Continue under Tier 3 Fractional CMO with AI Swarm ($15-25k/month, 6-month minimum, 1-2 days/wk embedded). Operator owns every artifact at every tier. Operator can in-house at any time.