Reading time: 3 minutes
Before I go into the best AI builds this week - I realised something.
People can't afford to just have things automated. They actually need to learn how to do it.
So we've launched the Automation Accelerator - A 4-week sprint to build + launch real automations while upskilling your team to do it themselves.
It's $5K one off. But because you're a special subscriber - you'll get 20% off. Just reply 'upskill'
1/ Jordan Government saved $10M by building an AI learning assistant
Tech Stack: Replit, OpenAI, Arabic NLP
They wanted to create an AI tutor for kids. Went to Scale AI who said they would do it for $10M. Someone non technical in Gov went to Replit and built it in 4 days.
Doesn't know how to code. Got all the countries curriculum, put it in AI and had a successful pilot.
Now they are scaling it to 2M+ students.
2/ Employee onboarding cut from 4 hours to seconds
TechStack: n8n, Google, Zoom, Slack, Microsoft
First impressions count. This is the desk of someone who joined Huel this week.
But their tech was set up the most the request was entered all via n8n:
🔧 User accounts set up
📭 Added to mailing lists
📍 Location configured
💬 Zoom, Slack, and Google profiles created
🔒 Secure passwords set
✅ Google Drive access granted
🏢 Office 365 licenses assigned
3/ / 11 year old builds Facebook for his school + they shut it down
Tech Stack: Lovable
This kid built an app where students can chat, post, comment, react, and even play word games and even keep streaks going.
He printed QR codes to invite his school mates. Regularly checked reports but then...
The school shut it down.
But if an 11 year old is building with AI. Why aren't you?
Read more here
4/ Mum of 3 build an app to create treasure hunts
Tech Stack: Replit, Notion, Google Maps, OpenAI
Ruby heard this idea on My First Million (great podcast if you don't know it)
She got building. Helps her go on adventures with her family and find treasure. Distributes via a dsicord community related to treasure hunting with 12k+ members
Here’s how she built it:
– Used Replit to code her app from scratch
– Connected Notion for note-taking, clues, and theories
– Integrated Google Maps to track routes and potential treasure sites
– Added a simple AI layer to help organize clues and summaries automatically
The result: cluemapper.com — like Notion + Google Maps for treasure hunters.
5/ AI-built Loan Officer Platform in 9 hours
Tech Stack: Lovable, Perplexity, Claude, Manus, Figma, Stripe
His friend had an idea to help loan officers manage their profiles and documents.
9 hours later, he had a full working prototype - built entirely through prompting in Lovable.
Here’s how it works:
– Users sign up and create profiles
– Documents are processed and classified automatically
– Loan officer profiles are auto-enriched with AI
– Payments handled through Stripe subscriptions
– Analytics baked in via Lovable
Everything — database, user management, edge functions, even the landing page — was done with Lovable Cloud.
Now his friend has a live MVP that would’ve cost ~$15,000 to build… for under $300 and one Saturday of “vibe coding.
Read more about it here
6/ LinkedIn Job scraper finds new jobs 24/7
Tech Stack: n8n, GPT-5, Apollo API, LinkedIn, Google Sheets
This guy was exhausted running his construction staffing agency - from manually scrolling LinkedIn every night — copying job posts, finding hiring managers, and updating spreadsheets.
So they built a set of AI agent automations in n8n that now run 24/7:
– Job Scraper: finds new LinkedIn job posts hourly
– Decision Maker Finder: identifies recruiters or HR leads
– Contact Enricher: pulls verified emails + company data via Apollo
– Research Agent: uses GPT-5 to generate personalised outreach based on each lead’s profile
He’s reaching hiring managers within the first hour a job post goes live — before 100+ applicants flood in.
Watch the step by step video here
Btw, I spoke at the Bubble conference last week about how we build successful apps with low code and AI. Watch it here
|
|
Harish Malhi
CEO, Goodspeed
|
