- Enquiry form → optional CRM
- Instant branded auto-reply
- Team alert on every lead
- Spam protection + fail-safe
I build owned, AI-automated operations stacks for service businesses — instant branded lead response, a 24/7 AI receptionist that captures and books, and content that publishes itself — running on infrastructure you own, not a $97/month (+$97 AI add-on) subscription you rent.
This is what happens when a customer submits an enquiry on a site I've built — every step automatic, every step logged.
A Sunshine Coast home-services business — my own, which I built this for first — was paying $97/month for an all-in-one SaaS (GoHighLevel), with its AI receptionist gated behind a further ~$97/month. Roughly $1,200–2,300+/year for a stack the business didn't own and would lose the day it stopped paying.
Recurring platform cost cut from $97/mo (+ optional $97/mo AI) to ~$0–15/mo — about $1,200–2,300+/year saved, with zero lock-in and full data ownership — while adding capabilities the SaaS never had. The whole stack is driven by one config file, so the same system re-stands for any service business that captures enquiries online.
Capture → instant branded reply → CRM → follow-up. The system above, fitted to your business.
Claude / GPT receptionists that answer 24/7, capture, book, and escalate hot / price / complaint to you — grounded in your real business knowledge, never quoting a firm price.
n8n, Make, Zapier — bookings, invoicing, notifications, CRM sync. The repetitive work, removed.
Forms, email, SMS, calendars wired into one pipeline that doesn't drop things.
Custom Vercel functions replacing monthly SaaS fees with infrastructure you own.
Deep GHL experience: automation inside it, or a clean migration off it. I've done both.
Most automation vendors pipe your customer data through whatever cloud tool is convenient. I engineer data boundaries: sensitive details (names, phones, medical / financial identifiers) are detected and redacted before anything leaves your infrastructure, and every byte that does leave is written to an audit log you can open.
For businesses that handle sensitive client information — clinics, brokers, legal, finance — this is the difference between "we use AI" and "we use AI responsibly." Ask me about local-first AI: models that run on hardware in your office, with the cloud used only for what it's genuinely better at.
One short call. You describe the outcome — faster replies, fewer missed leads, less manual admin — and I'll tell you exactly what I'd build, what it costs, and how long it takes.
alejandro.franco.portillo@gmail.com