automationscapeLet's talk
Menu

AI automation.
Built to work.

AI customer support, lead qualification and client acquisition systems. Built with n8n, Claude Code and Codex.

Let's build your workflow See the work
My n8n AI booking workflow, connecting the knowledge base, AI agent, service-area check and lead capture
AI booking workflown8n · OpenAI · Lead capture
400
QA hours saved per year with AI automation
10,000+
support tickets resolved at 92.6% CSAT
100+
app releases tested across desktop, mobile and browser

What I build.

AI customer support
& ticket resolution
For high-ticket SaaS companies.
AI lead qualification
& appointment booking
For high-ticket coaches and course creators.
Automated client acquisition
& appointment setting
For SEO and AEO agencies.

From chat
to booked call.

A cleaning-company booking assistant. Website, AI chat, lead capture and human handoff.

Tidewell cleaning-company demo website, designed and built by Ralph

From visitor to conversation.

A custom mobile-first website with an AI booking assistant.

Built as a demo project for a London cleaning company.

Open the full screenshot
The real n8n workflow connecting the knowledge base, OpenAI model, postcode check and lead capture

The work behind the chat.

n8n connects the knowledge base, OpenAI, postcode lookup and lead capture.

I tested 14 scenarios and fixed 12 issues.

Open the full screenshot
The Google Sheets knowledge base for the demo, including services, prices, policies and booking information

Answers from the source.

Prices live in a Google Sheet the team edits, so the bot never invents them.

The workflow reads the business information before it answers.

Open the full screenshot
The demo chatbot collecting Ralph’s test details and offering a quote-call booking link

One question at a time.

Checks the postcode, saves each lead to Google Sheets and sends a booking link.

When someone wants a person, it asks why and alerts the team on Slack.

Open the full screenshot

Demo project · n8n + OpenAI · Watch the video walkthrough

How I work.

  1. Map your process.

    Define exactly what the workflow should do.

  2. Test the failure paths.

    Work through messy and incomplete inputs.

  3. Build the handoff.

    Bring in a person for the cases AI can't resolve reliably.

n8n · Claude Code · Codex · Claude · OpenAI · Ollama
Zendesk · Intercom · Help Scout

Ralph Joel Bayot, AI automation developer

I'm Ralph.

Six years in crypto support and QA taught me to check the failure paths before calling anything finished.

I built an AI automation for a NYSE-listed crypto company that saved an estimated $20K a year. Now I build AI agents, chatbots and workflows for businesses.

What repeats?

Tell me what tools you use and where it currently breaks. We can start with one focused workflow and agree on how to measure it.

Let's talk about your workflow