Clean your CRM, flag deal risk

RevOps / sales-ops lead / VP sales / head of CS · Assemble in ~20 min; first hygiene scan + risk flags in <5 min once CRM export is ready.

GTMEdit flowon X◆ new (unverified) 86

step 1connector

Export CRM data

⚡ CRM (HubSpot / Salesforce / Sheet)

step 2ready

Scan for hygiene issues and duplicates

crm-hygiene-scanner3 alts

step 3partial

Apply CRM write-backs and corrections

crm-hygiene

step 4ready

Score and route inbound leads

revops

step 5ready

Process sales calls into CRM fields and follow-ups

post-call-followup

step 6ready

Flag deal risk and at-risk deals

deal-risk-review

step 7ready

Forecast and revenue model review

commercial-forecaster

Did this recipe work for you?

What you get

▸Hygiene report: duplicates, stale records, missing fields, and prioritized cleanup corrections
▸Scored and routed lead list with rep/queue assignments
▸Per-deal risk scorecard (red/yellow/green) with mitigation actions
▸Board-ready pipeline forecast (commit / best-case / pipe-only) with documented assumptions

tools: CRM subscription (existing) + optional Clay/Apollo for enrichment enrichment ($49+/mo if not already paying)

tokens: ~$2-6 per pipeline batch of 50-100 deals (hygiene + risk + forecast passes)

≈ $5-15/mo in tokens at weekly cadence; tool costs are additive to existing stack

What stays yours

· Review and approve write-back corrections before committing to CRM — do not auto-apply field changes without a human gate
· Validate routing assignments and override edge cases the scoring model cannot see (relationship context, exec sponsor dynamics)
· Take the flagged red-risk deals to a live rep conversation — risk mitigation actions are prescriptions, not autonomous outreach

The problem

When my pipeline data is messy and reps lose hours to admin, I want agents to score and route leads in real time, dedupe and update CRM records, auto-process every sales call into CRM fields/QA scores/follow-ups, and flag deal risk before deals go dark, so my revenue team spends time selling instead of on data entry.

Cybernetic layers

Memorybyo

Deal state, rep routing history, and prior risk flags need a persistent store — use a Google Sheet, Notion DB, or CRM itself. No harvested memory artifact covers this yet.

Background-schedulingpartial

For continuous hygiene (not just batch), the scan should run on a cron or webhook trigger when new records land. Claude Code scheduled tasks or a Zapier/Make webhook to re-export and re-scan can close this. Not yet a single library artifact.

Harnessready

Claude Code is the runtime. Drop each ready skill into ~/.claude/skills/ and invoke by name in sequence. Cowork terminal also works for one-shot sessions.

Click any step to see what fills it — and the ranked alternatives you can swap in.