CRM Data Decay Detector

AI audits Pipedrive contacts on a weekly schedule, detecting data decay across 5 categories and routing prioritized remediation briefs to your CRM.

AI audits Pipedrive contact records for data staleness — outdated job titles, email risk, company mismatches, ghost contacts — and delivers a prioritized remediation brief. 19-node swarm. $0.024/record. Runs on schedule, no manual effort.

Batch PipelinescheduleWeekly Cron01FetcherPipedrive — no LLM02Auditor5 categories — Opus 4.603Routerdecay tier?HIGHActivityMEDNoteLOWLog$0.024/record — single LLM callITP-measured across 20 records
CRM Data Decay Detector pipeline: batch audit with 3-tier decay routing to Pipedrive

One Auditor. Five Decay Categories. CRM Hygiene on Autopilot.

Step 1The Fetcher

Code Only

Pulls Pipedrive contacts sorted by last activity date. Filters to records stale beyond the configurable threshold (default 90 days). Zero LLM cost.

Step 2The Auditor

Tier 1 Reasoning

Scores each record across 5 decay categories: Title Staleness, Email Risk, Company Mismatch, Missing Critical Fields, and Ghost Contact. Single Opus 4.6 call per record.

Step 3The Router

IF Logic

Routes based on decay score and priority tier. HIGH → Pipedrive Activity (task to update). MEDIUM → Note for review. LOW → log only. Ghost Contact always escalates to HIGH.

What It Does NOT Do

×

Does not automatically update or delete CRM records — it detects decay and creates remediation briefs for human follow-up

×

Does not work with CRMs other than Pipedrive — no HubSpot, Salesforce, or custom CRM integration

×

Does not verify contact data against external sources — it scores decay signals from existing CRM data patterns

×

Does not process real-time events — it runs on a schedule and audits records in batch

The Complete Customer Success Bundle

9 files — everything you need to deploy the 19-node CRM Data Decay Detector pipeline.

crm_data_decay_detector_v1_0_0.jsonThe 19-node n8n workflow (batch-processing pipeline with 3-tier routing)
system_prompt_auditor.txtSystem prompt for the Auditor — 5-category decay taxonomy and scoring rubric
decay_score_rubric.mdComplete scoring rules (1-10 scale) with category weights and examples
error_handling_matrix.md24 failure modes, recovery paths, and dead letter handling
blueprint_dependency_matrix.mdPrerequisites, cost estimates, and n8n timeout configuration
itp_results.mdITP test results — 20 fixtures, 14/14 milestones, cost analysis
README.mdSetup guide — credentials, schedule config, batch size tuning, timeout settings
CHANGELOG.mdVersion history and release notes
system_prompt_fetcher.txtRecord selection logic and staleness filtering documentation

Tested. Measured. Documented.

Every metric is ITP-measured. The CRM Data Decay Detector audits stale Pipedrive contacts at $0.024/record with a single LLM call.

Workflow Nodes

19 specialized n8n nodes

Blueprint Quality Standard

BQS v2 — 12/12 PASS

Agent Architecture

1 (Auditor) + batch orchestration + Pipedrive sync

Required Credentials

Anthropic API + Pipedrive API

Bundle Contents

9 files

ITP Milestones

CDD-01 through CDD-08 + U-01 through U-06 — 14/14 PASS

Score Consistency

Variance 0 across 3 runs (threshold <= 2.0)

Cost per Record

$0.024 (ITP-measured blended average)

n8n Compatibility

Self-hosted

CRM Data Decay Detector v1.0.0 — Technical Reference
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Pipeline:     Schedule Trigger → Fetcher (Pipedrive API) → Staleness Filter → SplitInBatches → Auditor (Opus 4.6) → Router (IF) → Syncer (Pipedrive)
Topology:     Batch pipeline, single LLM call per record, 3-tier routing
Categories:   GHOST_CONTACT (30%) | TITLE_STALENESS (20%) | EMAIL_RISK (20%) | MISSING_CRITICAL (15%) | COMPANY_MISMATCH (15%)
Routing:      HIGH → Activity | MEDIUM → Note | LOW → Log | GHOST → always Activity
Node Count:   19 (Schedule 1, HTTP 4, Code 8, IF 3, SplitInBatches 1, Dead Letter 1, ITP Bypass 1)
LLM:          Claude Opus 4.6, temperature 0.2, max_tokens 2048
Cost:         $0.024/record blended | ~$1.21/week for 50-record batch
ITP:          20/20 (100%) — CDD-01 through CDD-08, U-01 through U-06
BQS:          12/12 PASS
Consistency:  Variance 0 across 3 runs (threshold ≤ 2.0)

CRM Data Decay Detector v1.0.0

$199

one-time purchase

What you get:

  • Production-ready 19-node n8n workflow — import and deploy
  • Scheduled batch processing — runs weekly, zero manual effort
  • 5-category decay scoring with weighted analysis
  • 3-tier routing: Activity (HIGH), Note (MEDIUM), Log (LOW)
  • Ghost Contact override — always escalates regardless of score
  • Configurable batch size, staleness threshold, and schedule cadence
  • Full ITP test results with 20 fixtures and cost analysis
  • BQS v2 certification (12/12 PASS)

Frequently Asked Questions

How does the scheduled batch processing work?+

The workflow runs on a configurable schedule (default: weekly). It fetches Pipedrive contacts sorted by last activity date, filters those stale beyond your threshold (default 90 days), and processes them in batches. No manual trigger needed.

What are the 5 decay categories?+

Ghost Contact (inactive 365+ days, no deals), Title Staleness (outdated roles, "Former/Ex-" prefixes), Email Risk (consumer domains on B2B records, role-based addresses), Missing Critical Fields (no phone, title, or company), and Company Mismatch (acquisitions, rebrands, defunct companies).

How does the 3-tier routing work?+

HIGH decay records get a Pipedrive Activity (a task assigned to update the record). MEDIUM records get a Note for human review. LOW records are logged only — no CRM action needed. Ghost Contacts always escalate to HIGH regardless of other scores.

Why is Ghost Contact treated differently?+

Ghost Contacts represent the highest-value cleanup targets — records with 365+ days of inactivity, zero deals, and minimal data. The asymmetric risk logic ensures these are always flagged for immediate action, never quietly logged.

How much does each record cost to process?+

ITP-measured: $0.024 per record blended average. A 50-record weekly batch costs approximately $1.21/week (~$5/month). Only one LLM call (the Auditor) per record.

What n8n timeout setting do I need?+

Set EXECUTIONS_TIMEOUT to 600 seconds minimum for the default 50-record batch. Each record takes 5-10 seconds for the LLM call plus CRM writes. The README includes exact configuration steps.

Can I customize the batch size and schedule?+

Yes. Batch size (default 50), staleness threshold (default 90 days), and schedule cadence (default weekly) are all configurable in the workflow. The README covers all tuning options.

What happens if a Pipedrive write fails mid-batch?+

CRM writes are non-blocking. Failed writes go to the Dead Letter Logger while the batch continues processing remaining records. No single record failure can stall the pipeline.

Is this the first batch-processing product in the lineup?+

Yes. All prior ForgeWorkflows products are event-triggered (webhooks or inbox triggers). CRM Data Decay Detector is the first scheduled batch workflow — it runs proactively rather than reacting to events.

What CRM does this work with?+

Pipedrive. The workflow reads contact records via the Pipedrive API and writes Activities and Notes back. It requires a Pipedrive API token with read/write access to persons, activities, and notes.

More Pipedrive Blueprints

Powered by Anthropic