Spec a CRM hygiene automation that keeps itself clean
Generates a maintenance-grade CRM hygiene spec: field standards, entry gates, scheduled sweeps with explicit detection logic, flag-queue ownership, and KPIs. Unlike a one-off cleanup checklist, it designs the system that prevents re-accumulation — the difference between mopping the floor and fixing the leak.
You are a CRM operations specialist who has cleaned up instances with 400,000 contacts and a decade of accumulated sludge, and your core belief is that one-time cleanups are theater — hygiene only holds when it's automated at the point of entry and swept on a schedule. Write a CRM hygiene automation spec for my instance. Output: 1. FIELD STANDARDS — for each core object (contacts, companies, deals): the fields that must be populated, allowed formats (picklist vs free text), and normalization rules (job title mapping, country codes, casing). 2. ENTRY GATES — validations and auto-enrichment that run when a record is created, so garbage never lands. 3. SCHEDULED SWEEPS — recurring jobs with cadence: dedupe pass, stale-deal flagger, bounced-email suppressor, orphaned-contact linker, title-change detector. For each: detection logic, auto-fix versus flag-for-human, and where flags surface. 4. OWNERSHIP — who reviews each flag queue and the weekly time budget it should take. 5. METRICS — 3-5 hygiene KPIs (duplicate rate, fill rate on critical fields, stale-deal percentage) with target thresholds. Rules: never spec silent auto-deletion of records — archive or flag, never destroy. Auto-fix only deterministic issues (formatting, casing, known mappings); anything requiring judgment gets a flag queue. Every sweep needs an owner or it will be ignored within a month. Before you write the spec, interview me. Ask ONE AT A TIME, waiting for my answer each time: 1. Which CRM, how many contacts and companies, and how old is the instance? 2. What are the three data problems that bite you most often? 3. Which fields does your team actually use for routing, reporting, or segmentation? 4. What enrichment tools do you have that could auto-fill gaps? 5. Who on the team can own weekly flag review, realistically? Then produce the spec, sized to my team — a two-person ops function gets a different spec than a solo founder.
How to use it
- 1
Copy the prompt into Claude, ChatGPT, or any LLM.
- 2
Answer the interview; question 3 is the anchor — hygiene effort should concentrate on fields that drive routing and reporting, not all fields equally.
- 3
Build the entry gates first (they stop the bleeding), then schedule the sweeps in your CRM or n8n.
- 4
Review the KPIs monthly and ask the chat to tighten whichever sweep isn't moving its number.
Best practices
Resist the urge to require many fields at entry — every required field trains reps to enter junk; require few, enrich the rest automatically.
Run the dedupe sweep in flag-only mode for two weeks before enabling any auto-merge.
Put the flag queues where people already work (Slack, CRM task views), not in a spreadsheet nobody opens.
Rerun this prompt after any migration or list import over a few thousand rows — bulk imports are where hygiene debt is born.
Example: what this looks like in practice
A RevOps manager inherits a six-year-old HubSpot instance with 180,000 contacts: three title formats, deals untouched for 200+ days still marked open, and a duplicate rate nobody has measured. She answers the interview naming routing fields (region, employee count, lifecycle stage) and Clay as her enrichment tool. The spec comes back with entry gates that auto-enrich company size on creation, five scheduled sweeps including a stale-deal flagger at 45 days idle, and a Monday flag queue budgeted at 30 minutes. After six weeks, critical-field fill rate climbs from 61% to 94%, and the sales team's segment reports stop needing manual cleanup before every QBR — which is the moment leadership stops questioning the CRM data in meetings.
Best fit
This prompt is one gear in a bigger machine. We orchestrate 20+ tools into outbound systems our clients own — and guarantee the results.
Apply for a Pilot Spot → →Frequently asked questions
Three layers: entry gates that validate and enrich records the moment they're created, scheduled sweeps that catch drift (duplicates, stale deals, bounces), and flag queues with a named owner for anything needing judgment. One-time cleanups without those layers decay back to baseline within a quarter — the system matters more than the scrub.
More automation & workflow design prompts
Build a system prompt for an AI cold email reply agent
Interviews you about your offer, booking flow, and voice, then generates a complete production system prompt for an AI reply agent — classification taxonomy, response rules, escalation conditions, and a strict JSON output contract an n8n or Zapier workflow can parse. You get the artifact teams pay thousands to have written, tuned to how you actually talk.
Turn a manual sales process into an automation blueprint
Converts a fuzzy 'we do this by hand every week' description into a structured blueprint with a detectable trigger, labeled steps, edge cases, human checkpoints, and failure alerts. It forces the design decisions people usually discover mid-build, so what you hand to n8n, Zapier, or an ops hire is assembly work, not archaeology.
Architect a Clay table before you burn the credits
Designs the Clay table before you build it: sources, column order, run conditions, waterfalls, outputs, and a credit estimate. Column ordering and gating are where most teams silently overspend, so the prompt enforces filter-first architecture and makes every paid column defend its run condition — typically cutting credit burn 40-70% versus enrich-everything builds.