※ AGENCY RECEIPT
Your agency,
itemized.
A weekly receipt for AI-automation agencies. What worked. What broke. What to fix this week. Built for operators running n8n, Make, and GHL on behalf of clients.
──── ITEMIZED ────
- One workflow you ran last week, audited line-by-line — qty 1
- The 3 things to fix this week, ranked by hours saved — qty 3
- One pattern from another agency you can steal — qty 1
- A literal receipt PDF, in your inbox, every Monday — qty 1/wk
TOTAL $0.00 — first 5 agencies
Bring one workflow that broke this week. I tell you what I'd fix, in what order, and what it'd cost in hours. No pitch. No deck. No follow-up sequence.
──── SAMPLE: RECEIPT #001 ────
RECEIPT #001 2026-04-28 — 2026-05-04
Agency
Northbeam Automation Co. (8 active client workflows)
Audited workflow
n8n → GHL: Lead handoff with UTM enrichment
What broke
- • UTM params dropped on 41% of leads (HTTP node not forwarding query string)
- • Retry logic = none. 14 leads silently lost on 502 from GHL.
- • No logging. Could not reproduce the failures without DM-ing the client.
Fix order this week
- 1. Pass through query string on HTTP node — est. 12 min
- 2. Add 3-attempt exponential retry on GHL POST — est. 18 min
- 3. Pipe failures to a Slack channel + Postgres log — est. 45 min
Steal this from another agency
Wrap every external POST in a sub-workflow that returns a typed result
("ok" | "retryable" | "fatal"). One agency cut their Slack incidents 70%
doing only this.
Hours saved if you fix all 3 ~11 hrs / week