VisionQuantech n8n Automation Pack: 20 Deployable Workflows for Local Businesses (Open Source)

TL;DR: An open-source pack of 20 importable n8n workflow templates plus Python builder scripts that generate them — lead responders with HOT/WARM/COLD scoring, appointment bookers with 24-hour reminders, AI calling and WhatsApp agents, review and SEO agents, and a follow-up engine. The v1 core (three workflows) was built and tested on n8n 1.83.2; setup needs zero credentials for the demo and a single Google OAuth credential for the production workflows.

The problem: leads die in the gap between enquiry and response

A clinic's website form gets an enquiry at 9pm. Nobody reads it until morning. By then the patient has booked with a competitor. The same story plays out for hotels, salons, real-estate agencies and consultants: the business is closed, busy, or on leave, and the lead goes cold in hours.

The fix is unglamorous automation: respond instantly, score the lead, book the appointment, log everything, remind before the slot, and follow up when nobody replies. n8n is the natural platform for it — self-hostable, visual, and cheap. What's usually missing is a ready-made, sellable pack of workflows a small business (or the agency serving it) can import, connect, and activate without weeks of node-wiring. That's what this repository is.

What's in the pack

The repository contains 20 workflow JSON files under workflows/, four Python builder scripts (build_pitch_workflows.py, build_pitch2_workflows.py, build_pitch3_workflows.py, build_batch3_zips.py) that generate workflows programmatically, and documentation: README.md, SETUP.md (a real step-by-step setup guide), SALES.md (one-page sales pitches per workflow, written for clinics, hotels and real estate), BACKUP_STATE.md, and protocol.md.

The tested v1 core: three workflows that actually ran

The README is explicit that three workflows were built and tested on n8n 1.83.2:

1. Instant Lead Responder — Demo (vq-lead-responder-demo.json): a website-form lead arrives via POST webhook, gets auto-qualified as HOT/WARM/COLD by scoring keywords plus contact completeness, and receives a personalised instant reply as JSON. Needs nothing — zero credentials, zero API keys. It exists to be demonstrated live on a sales call.
2. Instant Lead Responder — Pro (vq-lead-responder-pro.json): the same qualification engine, plus the instant reply goes out by Gmail and every lead is logged as a row in a Google Sheet — name, phone, email, service, score, tier, reply sent, next action. Needs one Google OAuth credential and a blank Sheet with a tab named Leads.
3. Appointment Booker + Reminders (vq-booking-plus-reminders.json): a booking form (name, phone, email, service, date, time) creates a Google Calendar event, sends a confirmation email, responds instantly, then waits until 24 hours before the slot and sends a reminder email — using n8n's Wait node, which resumes correctly after restarts.

The expanded suite

Beyond the v1 core, the pack includes agents for the channels and jobs local businesses actually need:

Under the hood: what one workflow actually does

I inspected vq-ai-enquiry-agent.json node by node rather than trusting the filename. It is a compact four-node pipeline, and the interesting engineering sits in the Code node:

Enquiry Webhook (n8n-nodes-base.webhook)
   → Detect Intent (n8n-nodes-base.code)
   → Needs Human? (n8n-nodes-base.if)
   → Send Reply (n8n-nodes-base.respondToWebhook)

The Detect Intent node is plain JavaScript over the incoming JSON. It normalises the payload (accepting either a wrapped body object or a flat object), then matches the message against five keyword lists — booking, pricing, info, complaint, and urgency words like "urgent", "asap", "today", "emergency", "call me". Intent resolution has a deliberate priority order: complaint beats booking beats pricing beats info, and urgency is OR-ed with complaint to set a needsHuman flag. Replies are channel-aware: if the channel is WhatsApp or SMS, the response is the short variant; web gets the longer one. The Needs Human? IF node then routes complaints and urgent messages to a human path instead of auto-replying — a small but important safety property: the agent escalates anger, it doesn't argue with it.

Two more workflows deserve a mention for what they reveal about the pack's commercial thinking. vq-ad-lead-responder.json exists because paid-ad leads decay faster than organic ones — every minute of latency burns ad spend — so it gets its own responder rather than sharing the generic one. vq-afterhours-catcher.json is the night-shift safety net: when the business is closed, enquiries are captured, acknowledged, and queued for the morning instead of evaporating. And the documentation goes beyond setup: protocol.md records the operating protocol for running these workflows as a service, while BACKUP_STATE.md tracks instance state — the unglamorous bookkeeping that separates a hobby deployment from one you can sell and support.

The builder scripts take this further: instead of hand-wiring every workflow in the n8n editor, the Python generators (build_pitch_workflows.py and siblings) produce the workflow JSON programmatically — the same pattern used to build out the pitch and batch suites. Generated-but-readable beats hand-drawn when you're shipping twenty of these.

How to use it

1. Import (2 minutes per workflow): in the n8n editor, Workflows → ⋯ → Import from File → select the .json. Save it.
2. Connect Google once (5 minutes, Pro and Booking workflows only): create a Gmail OAuth2 credential in n8n with the Google account that will send mail, then attach the same credential to the Sheet-logging node (pointing at a blank Sheet with a Leads tab) and the Calendar node. The Demo workflow skips this entirely.
3. Activate and get the live URL: toggle the workflow Active, open the Webhook node, and copy the Production URL (.../webhook/lead-in, .../webhook/lead-in-pro, .../webhook/booking-in). Point the client's website form at it with a POST JSON payload of {name, phone, email, service, message} — or {name, phone, email, service, date, time} for bookings.
4. Test in 60 seconds with curl against the webhook, e.g.:

   curl -X POST https://<your-n8n>/webhook/lead-in \
     -H "Content-Type: application/json" \
     -d '{"name":"Test Patient","phone":"+919999988888","email":"[email protected]",
          "service":"dental implant","message":"I need an urgent appointment tomorrow"}'
   # Expect: {"tier":"HOT","score":>=70,"reply":"Hi Test Patient! ...", ...}

The repo's suggested pricing anchors the pack as a product, not just code: free live demo, a one-time setup fee in the same slabs as the AI-agent pitches, and an optional monthly care plan for monitoring and tweaks. SALES.md even scripts the objection handling — the "we already have a receptionist" answer is that the workflow covers the gaps (leave, busy, after-hours), not the person.

Honest status and roadmap

Working. The v1 trio was built and tested on a live n8n 1.83.2 instance, the setup guide documents real webhook URLs and real payload contracts, and the inspected Code node does exactly what the docs claim. The README also carries honest caveats instead of hiding them: n8n shows a critical-update banner (updating needs a redeploy, not yet done); the free-tier host sleeps when idle so the first webhook hit after sleep takes ~30–60 seconds to wake (production clients should use an always-on host); and a second n8n instance is currently locked out, so everything runs on the working one.

What it isn't: the expanded suite beyond the v1 trio is import-ready rather than individually test-documented, and every production deployment still needs the client's own Google OAuth credential wired in — there's no magic multi-tenant provisioning here. Sensible next steps: per-client credential templates, a packaged Docker deployment so buyers skip the n8n learning curve, and converting the Python builders into a proper workflow SDK with versioned releases.

The bottom line

This is automation as a product: twenty workflows, a setup guide that reads like it was written by someone who actually deployed them, sales scripts matched to the buyer, and honest notes about hosting sleep, update banners, and locked instances. The v1 core is tested and live; the rest is a genuine head start for any agency or freelancer selling automation to local businesses.

Repository: github.com/Shivay00001/n8n-automations · Want this deployed for your business? VisionQuantech sets up lead responders, booking engines and follow-up automation for clinics, hotels, real estate and more. Contact: [email protected]