Lead Leak Check — Zapier, Make & n8n Lead Pipeline Audit

Your Zapier, Make or n8n runs say Success, but some leads still never hear back. Nobody owned the reply, a filter stopped on an empty field, the workflow was off for a day, or an AI draft sat unsent.

Lead Leak Check audits your lead pipeline / lead-to-CRM workflow, not your website. You paste a plain-English description, a JSON config, or a run-history CSV export; it returns a 0–100 leak-risk score with ranked findings, and each finding says why leads leak there and gives a concrete fix. Built for agencies and automation freelancers auditing a client's lead routing, sales ops checking speed-to-lead safeguards, and founders running leads through Zapier/Make/n8n into Sheets, HubSpot or Slack.

• No logins, no scraping, no API keys. Everything runs on 15 offline rules over what you paste. Your data is only used for this run.
• It does not crawl or test websites or contact forms. It only knows what you describe or export.
• Unstated safeguards are reported as verify, never as safe, so a vague description never looks like a clean bill of health.
• $0.05 per report, the demo report for an empty input is not charged as a report (details under Pricing).
• The same rules are open source as an MCP server: lead-leak-check-mcp.

How to use it

1. Click Try for free / Start. The input is pre-filled with an example pipeline, so the first run works as-is.
2. Replace the example with your own pipeline: a few sentences, a JSON config, or a run-history CSV export.
3. Open the Output tab: one row per report with the risk score and ranked findings, plus the full Markdown report (REPORT.md).

Runs finish in a few seconds and use the smallest memory size.

Input

Fill in any one or more of these fields. Each one produces one report, which is one dataset item. If you start a run with no input at all (for example {} through the API, a schedule or an integration), the Actor analyses a built-in sample pipeline and returns one free demo report marked "inputType": "sample", so you can see the output format.

Field	What to paste
pipelineDescription	Plain English: "Webflow contact form into HubSpot via Zapier. Nobody is assigned. No error alerts. Nobody works on weekends."
pipelineConfig	Structured JSON for exact results (shape below)
runHistoryCsv	Run-history export with a status column and, optionally, a timestamp column
expectedIntervalMinutes	Optional. Expected minutes between runs, used for the CSV gap check

{
  "name": "Typeform -> Google Sheets -> Slack #sales",
  "sources": [{ "type": "web_form", "required_contact_field": false }, { "type": "whatsapp", "logged_to_storage": false }],
  "storage": { "type": "google_sheets", "dedupe": false },
  "routing": { "owner_assignment": "none", "after_hours_coverage": false },
  "response": { "sla_minutes": 30, "auto_ack": true, "ai_drafts": true, "human_send_step": false },
  "monitoring": { "error_alerts": false, "unanswered_lead_alert": false, "run_history_review": "never" },
  "automation": { "platform": "zapier", "plan": "free" },
  "timezone": { "business_tz": "America/New_York", "automation_tz": "UTC" },
  "follow_up": { "sequence": false }
}

Output

One dataset item per report. This is the real (shortened) result for the pre-filled example, "Leads come from a Webflow contact form into HubSpot via Zapier. Nobody is assigned; new leads just post to Slack #sales. No error alerts if the Zap fails. Nobody works on weekends, so Friday-night leads wait until Monday. No follow-ups after the first reply." (4 rules fire: 2 high, 1 medium, 1 low; 8 safeguards not mentioned are listed as verify):

{
  "inputType": "description",
  "pipeline": "described pipeline",
  "risk_score": 82,
  "counts": {
    "high": 2,
    "medium": 1,
    "low": 1,
    "verify": 8
  },
  "findings": [
    {
      "id": "no-owner",
      "severity": "high",
      "title": "Nobody owns the first reply",
      "why": "Leads posted to a shared sheet or channel are seen by everyone and answered by no one.",
      "fix": "Assign an owner per lead (round robin or fixed) at the moment it is created, and name them in the notification.",
      "evidence": ""
    },
    {
      "id": "no-error-alert",
      "severity": "high",
      "title": "Failed or disabled automations are silent",
      "why": "Expired connections, plan limits or a teammate's edit can switch a workflow off with no notification.",
      "fix": "Turn on error notifications (Zapier/Make/n8n error workflow) to a channel someone watches, and alert when no run happened in the expected window.",
      "evidence": ""
    },
    {
      "id": "no-after-hours",
      "severity": "medium",
      "title": "No after-hours or weekend coverage",
      "why": "Leads that arrive at night or on weekends wait until the next business day, often past the buyer's patience.",
      "fix": "Send an instant acknowledgement with the next reply time, and escalate weekend leads to an on-call person.",
      "evidence": ""
    }
  ],
  "notes": [
    "Parsed from free text with keyword rules; send a JSON config (see check_pipeline) for exact results.",
    "8 safeguard(s) were not described; they are listed as 'verify'. Fill them in for a sharper result."
  ],
  "report_markdown": "# Lead leak risk: ..."
}

Field	Meaning
inputType	description, config, run_history, or sample (built-in demo for an empty input)
risk_score	0 = no known risk, 100 = leaking everywhere
counts	Findings per severity: high, medium, low, verify (not stated, please check)
findings	Ranked high to verify; each has why and a concrete fix
report_markdown	The same report as Markdown, ready to paste into a doc or ticket

All reports of the run are also combined into one Markdown file, REPORT.md, in the run's key-value store. Both are linked on the Output tab and available through the API (/datasets/{id}/items and /key-value-stores/{id}/records/REPORT.md).

What it checks

Owner assignment, unanswered-lead (SLA) alerts, error alerts, response-time target, system of record, chat leads (WhatsApp/DM) not logged, silent filters on empty fields, unsent AI drafts, after-hours coverage, instant acknowledgement, deduplication, timezone mismatch, free-plan limits, follow-up sequence and run-history review. From a run-history CSV it flags failed and halted runs, quietly filtered runs, long gaps with no runs and a stale last run.

Pricing

Pay per report: $0.05 per report (one dataset item), plus Apify's standard Actor-start event of $0.00005 per GB of memory per run. There is no subscription. A run with all three fields filled in produces three reports. Empty or invalid input is never charged, and the demo report for an empty input is not charged as a report (only the tiny Actor-start event applies). If you set a maximum charge per run, the Actor stops before a report it cannot charge.

FAQ

Does it scan my website or test my contact form? No. It does not fetch any URL. Describe the form and where its leads go, and the rules check the pipeline behind it.

Does it log in to my Zapier, Make or n8n account? No. It only reads what you paste. Nothing is fetched or stored outside your own Apify run.

Why do I get verify findings? The description did not mention that safeguard. The Actor will not assume it exists. State it (for example "errors alert #ops") or use the JSON config to clear it.

Can I call it from Zapier, Make, n8n or an AI agent? Yes. Use the Apify integration or the API with a JSON input, then read the dataset items or REPORT.md.

Run locally

pip install -r requirements.txt
mkdir -p storage/key_value_stores/default
echo '{"pipelineDescription": "WhatsApp leads land in HubSpot via Make.com. Nobody is assigned."}' > storage/key_value_stores/default/INPUT.json
APIFY_LOCAL_STORAGE_DIR=./storage python -m src
pip install pytest && python -m pytest -q tests

Or apify run with the Apify CLI.

Made by Paopao Studio. MIT license.