Skip to main content Founding cohort open 3 free builds left for our first clients Claim one
Build in public

Our first audit
was our own
business.

Before we sold this audit to anyone else, we ran it on Seasonal Services — our own Christmas-lights company. This page is the receipt. What we mapped, what we built, what we expect, and (after October 2026) what actually happened.

Live: season 1 running. Numbers publish Oct 2026.
What we automated

Four automations, running live.

Each one came out of the audit of Seasonal Services. Each one has a specific hypothesis we’re measuring against. Results in October.

  • 01 Live

    Lead response

    Stack: n8n · OpenAI · Twilio

    AI triages form, call, and text submissions and replies in under 4 minutes. Escalates to me only if the lead is hot or off-script.

    Hypothesis

    Cut missed-lead revenue leak by 60–80%.

  • 02 Live

    Review request flow

    Stack: n8n · Jobber webhook · SMS

    Triggered on job completion. Branded message, one auto-follow-up, then stops. Uses first name and install location.

    Hypothesis

    Double the review-request-to-review conversion vs. manual asks.

  • 03 Live

    Quote follow-up

    Stack: n8n · OpenAI · Email

    3-touch sequence, AI-drafted per lead with their install details. Hands off to me the moment they reply.

    Hypothesis

    Recover 15–25% of quotes that used to go silent.

  • 04 Live

    Data sync

    Stack: n8n · Jobber · QuickBooks

    Two-way sync for jobs and invoices. No more copy-paste on Sunday nights.

    Hypothesis

    ~2 hours a week of my life back.

Timeline

From mapping to measurement.

  1. MAP
    Aug 2025

    Drew the whole business on one canvas.

    Every lead source, every handoff, every tool. Jobber + QuickBooks + a shared inbox + a lot of text threads. Twelve distinct workflows. I didn't know there were twelve until I drew them.

  2. AUDIT
    Aug 2025

    Scored each workflow on three axes.

    Hours it cost me per month, dollars it leaked per month, how hard it would be to automate. The biggest leaks weren't the ones I'd been worrying about. The actual number-one leak was after-hours leads — people filling out the form at 9 PM who I couldn't reach until I came down off a ladder the next afternoon.

  3. PLAN
    Sep 2025

    Picked four automations. Killed two.

    Six candidates made it through the audit. I killed two that looked cool but wouldn't pay back in six months. The other four: lead response, review request, quote follow-up, and data sync between Jobber and QuickBooks.

  4. BUILD
    Oct 2025

    Shipped all four. Ugly first, pretty later.

    Built in n8n, with OpenAI for the conversational pieces. Rough edges at first, then refined over a few weekends. Everything is running live in Seasonal Services right now.

  5. RUN
    Nov 2025 — Oct 2026

    First full Christmas-light season running on the new stack.

    The season is happening as you read this (or has just finished, depending on when you're reading). I'm tracking: leads captured, response times, reviews generated, hours saved, and dollars recovered.

  6. PUBLISH
    Oct 2026

    Numbers go public. Good or bad.

    I'll update this page with the full before/after. If the automations worked, you'll see it. If they didn't, you'll see that too. That's the whole point of doing it in public.

Accountability

If the numbers are bad, we'll tell you the numbers are bad.

In October 2026, this page updates with the full before/after. Leads captured, response times, reviews generated, hours saved, revenue recovered. No cherry-picking. If a hypothesis didn't pay out, we'll explain why and what we learned.

Book a Free Audit Call