The Idea: In April 2023, two McGill University classmates - Max Brodeur-Urbas and Rahul Behal - started building a side project in a Vancouver bedroom. They were tinkering with Auto-GPT, the open-source autonomous AI agent that had just gone viral, trying to wrap it in something a normal person could actually use. The wrapper turned into a proof of concept. The proof of concept turned into a startup. They called it AgentHub. Within months they'd joined Y Combinator - one of the few Canadian companies in the Winter 2024 cohort and shipped a no-code workflow automation platform where anyone can drag modular AI components onto a canvas and wire them together. This was the beginning of Gumloop.
The Execution:
The lesson? Max and Rahul didn't set out to build a $70M company. They set out to make Auto-GPT usable by a normal person. The gap between "this technology is technically possible" and "this technology is actually accessible to a non-developer" turns out to be worth a lot of money. The founders who win automation aren't always the ones who build the most powerful tool - they're the ones who make the powerful tool feel approachable. When the entire workforce becomes your potential user, the addressable market gets very interesting very fast.