Jonathan Bayo

Case studies

I build automation systems that do the work people shouldn't have to.

I'm Jonathan Bayo. I build lead research, outreach, CRM and reporting systems in n8n, Python and Claude for founders and small teams. 14 of them are below, with how each one runs and what came out of it.

Usual stack
n8n, Python, Claude, Supabase, Apify
Typical build
one to two weeks
Where things run
your own accounts and servers
Right now
taking new builds

The rest, one by one

Live running for my own business right now Client built for a client and handed over Built a working system that proves the pattern, not deployed for a client yet

OperationsClient

Workflow watchdog and uptime monitor

A daily n8n job that checks every other workflow, tells a stall from a weekly schedule, re-fires what it can, flags what needs money, and only emails a person when it can't fix the problem itself.

n8n · n8n public API · Supabase · Microsoft Graph mail · GitHub Actions

Dailychecks 28 workflows
5 minuptime probe
Read the case study
AI systemsClient

Personalized video generator

Picks the highest priority leads from Supabase, has Claude write a 45 second script, sends it to Visla to render and saves the video link back on the lead.

n8n · Claude API · Visla API · Supabase · HMAC auth

5 mincycle time
0manual steps
Read the case study
IntelligenceClient

Company intelligence monitor

Every Monday it reads a list of target companies, searches for fresh signals like funding, hiring and leadership changes, scores them and pushes the hot ones into the CRM.

n8n · Airtable · Tavily search · Jina search · GoHighLevel

Mon 06:00weekly run
200+companies watched
Read the case study
IntelligenceClient

CRM lead enrichment pipeline

Takes a scored lead from the monitor, has Claude write a two sentence problem statement for the company, builds the full CRM record and pushes it to GoHighLevel. About three seconds a lead.

n8n · Claude API · GoHighLevel · Webhooks · JavaScript code nodes

~3 sper lead
0manual entry
Read the case study
OutreachBuilt

Google Maps lead discovery

Searches Google Maps for a business type in a city, scrapes each website for contact details, has Claude pull out what the business does and writes a clean lead list to Google Sheets.

n8n · Apify (Google Places) · Python · Claude API · Google Sheets

500+leads a run
~80%email find rate
Read the case study
AI systemsClient

Internal knowledge assistant

Staff ask a plain question about company documents, policies and contracts and get an answer with the source cited. Runs on the company's own servers.

Python (FastAPI) · Weaviate · LangChain · Claude API · Docker

~2 sper answer
40%fewer tickets, reported
Read the case study
IntelligenceClient

Recruitment enrichment engine

Scrapes candidate profiles for a role, enriches them with public data, scores fit against the job with AI and pushes only the qualified ones into Airtable with a summary for the recruiter.

Make.com · Apify · OpenAI API · Airtable · Slack

500+candidates a day
40%reach a human
Read the case study
OperationsClient

Zero touch client onboarding

A signed form kicks off the contract, the e-signature, the project folder, the Slack channel, the CRM deal and the welcome email. About 60 seconds, no hand offs.

n8n · Google Forms · DocuSign · Google Drive · Slack

~60 send to end
~8 hsaved per client
Read the case study
OperationsBuilt

Investor report automation

On the first of the month it pulls portfolio numbers from Sheets, has Claude write the month's narrative, builds each investor's report with their own allocation and sends it.

n8n · Google Sheets · Claude API · Gmail

1stof every month
~20 hsaved a month
Read the case study
OutreachBuilt

Event follow up engine

After an event, every attendee gets a follow up within 24 hours that refers to the sessions they actually attended. The most engaged ones get flagged for a personal call.

n8n · Claude API · Gmail · Airtable · Webhooks

24 hfollow up window
300+attendees an event
Read the case study

About

I've spent the past year as the automation and back end engineer for a B2B sales intelligence company, building the research pipeline, the CRM tooling and the multi-tenant product you see above. Alongside that I take on builds for founders and small teams: outreach, lead research, onboarding, reporting.

I work in n8n and Python, use Claude where a step needs judgment, and keep everything in your own accounts so nothing depends on me being around.

How I work

We start with what's taking too long. I map the steps, pick the smallest system that removes them and build it, with Claude where a step needs judgment.

Everything logs its own runs, never sends the same thing twice and fails loudly instead of quietly. You get the workflows, the credentials in your accounts and a short doc on how it runs and what to do when it stops. Most builds take a week or two. If it's going to take longer I'll say so before we start.

Want something like this built?

Tell me what's taking too long. Most things can be built in a week or two.