π A deal gets marked "Won" in HubSpot.
Less than a minute later...
β
The correct invoice is generated
β
The customer receives it by email
β
The invoice is logged automatically
β
The team gets notified on Slack
Nobody creates or sends an invoice manually.
Here's how I built this workflow using n8n + HubSpot. π§΅π
One thing I always think about while building automations is...
"What happens when something goes wrong?"
For example, not every HubSpot deal is linked to a contact.
Instead of letting the workflow fail, it detects the issue and alerts the team immediately.
Production-ready automation isn't just about the good path alone. It's about handling the edge cases too.
1
26
This is Automation 1 of 2 for working with HubSpot CRM with automation.
Built with:
β‘ n8n
π’ HubSpot
π Google Sheets
π Google Drive
π§ Gmail
I document every AI and n8n automation I build.
Follow for more real-world automation builds.
#n8n #AIAutomation #WorkflowAutomation #HubSpot #BusinessAutomation #NoCode #Automation #CRM
48
Webhooks explained, but let's make it practical π
You don't need to build or host your own backend to catch webhooks, n8n and Make.com already run one for you.
βοΈ In n8n:
β
Add a Webhook node, it instantly generates a URL.
β
Paste that URL into the source app (Stripe, Shopify, Typeform, etc.) as its webhook endpoint.
β
Every event now lands directly in your workflow.
β
Chain it to any action node, update a sheet, send a Slack message, create a record.
π In Make.com:
β
Add a Webhook module as your scenario's trigger.
β
Click "Add hook" to generate a custom URL.
β
Paste it into the external service's webhook settings.
β
Run once to capture sample data, then map fields into your next modules.
β
Activate the scenario, it now runs automatically on every event.
No servers. No custom code. Just register the URL, connect the nodes, and let the automation run itself.
#webhooks #n8n #makecom #automation #api #nocode #devtips
My automation almost told a client their delivered package was still on a truck.
A "delivered" email came in late, and my flow overwrote the record β bumped it back to "in transit."
Building the automation was easy. Building it so it can't confidently be wrong is the actual work.
Lessons π§΅
6/ Never let it go quiet.
An error you can see. The scary one is the shipment that silently gets skipped until the customer calls.
If your flow can go quiet on bad input, fix that first.
1
45
7/ What it does now:
email β reads it (scanned PDFs too) β pulls tracking + carrier + status β updates the board. Creates new, updates existing, never moves backwards.
n8n + AI agent + Airtable.
Phase 1. Full build on my channel π
youtu.be/6LCUC3B3wsE
1
43
Alao Abiodun retweeted
π€ Built a Voice AI Agent using ElevenLabs + n8n.
This AI can:
β
Handle real estate inquiries
β
Qualify leads
β
Check appointment availability in real time
β
Suggest alternative time slots automatically
AI agents are changing how businesses manage customer interactions
Alao Abiodun retweeted
Part 2 π
After checking availability, the AI automatically schedules the appointment and updates the calendar.
Built with ElevenLabs + n8n.
Watch the quoted post first for Part 1 π
#AIAgent #VoiceAI #n8n #ElevenLabs
Alao Abiodun retweeted
Part 3 π
Not every real estate prospect is ready to book an appointment.
This AI captures their contact details, property interests, and requirements for future follow-up, helping businesses build a stronger pipeline and reduce lost opportunities.
#AIAgent #VoiceAI #ElevenLab
Part 2 π
After checking availability, the AI automatically schedules the appointment and updates the calendar.
Built with ElevenLabs + n8n.
Watch the quoted post first for Part 1 π
#AIAgent #VoiceAI #n8n #ElevenLabs
Alao Abiodun retweeted
The whole thing in one line:
Documents β chunks β vectors β stored. Then: question β search β retrieve β grounded answer.
This is the exact pattern I build for clients. And now you know how RAG actually works.
What would you point your AI's knowledge at first?
Alao Abiodun retweeted
Ever worked on RAG development?
Think of it like giving the AI an open-book exam instead of forcing it to answer from memory.
Most people assume AI just knows everything and answers from memory. That's exactly why it hallucinates.




