AI Automation · SaaS · Conversational AI

I live inside AI systems, automations, products. This is all I do.

I am Bishal Paul. I ship production grade AI automation systems, SaaS products, and conversational AI agents for teams that need them running on Monday, not in a demo.

Book a call See what I have shipped

◉ AI Automation◉ n8n, Supabase◉ AI Products, SaaS◉ Conversational AI◉ Claude API

Bishal Paul · United Kingdom10 yrs shipping · since 2016 →

$ ./intro.sh

Fig. 00 · Live shell

~/bishal.paul, zsh

bishal@paul $

How I work

Fig. 01 · Engagement flow

Four stages ·
zero shortcuts

Every engagement moves through the same four stages. Discover the real problem, design the system before writing code, build against production failure modes from day one, and hand off something the team can operate without me in the room.

What I build

Five practice areas, one operating principle: ship systems that survive production.

Ordered by how I actually spend my week. Each links to a longer view of the work and where I plug in.

01 / 04

AI Automation Systems

Self hosted n8n and Supabase systems that quietly run the parts of a business no one wants to touch again.

n8n
Supabase
DigitalOcean
Docker Compose
Webhooks

02 / 04

AI Product & SaaS Development

Shipping AI powered products from prompt to pricing page, with a founder's judgement on what to build and what to skip.

Claude API
Next.js
Supabase
Better Auth
MongoDB

03 / 04

Conversational AI & Voice Agents

Production voice and conversational agents that handle real inbound and outbound conversations, from routing to payment capture.

ElevenLabs
Twilio
Vonage
n8n
Supabase

04 / 04

AI Support & Ops Systems

Ticket triage, agentic ops, and internal tools that let a small team behave like a much larger one.

Freshdesk
Claude API
n8n
Supabase

See all five pillars on the About page →

Choose a problem

Pick a real problem. See the system I would ship.

Every tab below is a system I have taken to production. Choose one to reveal the stack, the pipeline, and the guardrails.

System · 01

Assisted operator system. AI drafts, humans send.

Every reply traceable to a prompt version. Escalations flagged automatically. No silent auto send.

FreshdeskClaude APIn8nSupabase

./pipeline

01New ticket→
02Intent classify→
03KB lookup→
04Claude draft→
05Human review→
06Freshdesk send

Systems blueprint

Three systems I keep shipping. Look inside each one.

Toggle a blueprint. See the data flow, the stack, and the outcome the system is designed to produce.

blueprint.render

AI Automation System

Self hosted n8n plus Supabase, running the parts of the business no one wants to touch again.

Stack

n8n · Supabase · Docker · DigitalOcean · Webhooks · Cron

Outcome

Workflows a small team can operate without a platform team behind them.

What changes

Manual hours removed: 20 to 60 per week
Time to first live workflow: 2 to 4 weeks
Runs per month at steady state: 10k to 500k

Typical use cases

◉Lead routing and enrichment from forms, LinkedIn, and inbound email
◉Finance and billing reconciliation across Stripe, banks, and ledgers
◉Ops handoffs between CRM, project tools, and internal dashboards
◉Recurring reports, exports, and scheduled data syncs

Also known as

n8n automation
self hosted workflow automation
AI workflow automation agency
Supabase automation stack

// data flow, left to right

01◉
Sources
FTP, webhooks, forms, email
02◉
n8n orchestration
Branching, retries, cron
03◉
LLM reasoning
Claude, structured outputs
04◉
Supabase
Idempotent upserts, audit log
05◉
Systems of record
CRM, billing, ops tools

Path: src→n8n · n8n→llm · llm→n8n · n8n→db · db→out

Featured solutions

Systems I have shipped, not services rendered.

Anonymized at the industry level. Every solution below runs in production today.

Telecom Sector

Voice AI for Outbound Debt Collection

An outbound voice agent that runs collections calls at scale, detects voicemail, falls back to SMS, and captures phone payments without a human on the line.