(1) you have to decide if are you "adopting" AI, or rebuilding for AI?

→ We have a new team of AI teammates, and they have to work on their tech stack that will continue to change... (I know how it sounds, but trust me, it's 100% like that)

(2) our AI teammates don't code like us, so we need to find a better way to code with them and manage their work.

→ When iterations are that fast, we needed the flexibility to run multiple agents and use Slack to manage our new team

→ New workflow to code with Cursor background agents → Vercel → GitHub → Live (from text to live feature)

(3) (biggest surprise for me, at least) we are killing our existing hosting and DB.

Our new AI teammates need their own hosting and DB for AI. If in the old world companies are running on Google or AWS and it's a nightmare to switch, we needed to run on an AI-native DB with a switch overnight.

→ Hello new hosting and DB for AI teammates (nice to also save $100K :)

(4) New native AI Web app.

I still remember that a couple of months ago, we were still considering using a Material UI template. It's alienating your AI team and putting constraints on your product.

→ Completely new web app for our products built in 2 weeks and better than anything we had before.

(5) New sense of speed.

This is key. The humans in the team are really experienced developers and PM’s, but can they completely change the way they work for this new type of speed?

→ mentality shift is the hardest. The change didn't happen overnight. It took us 4-5 weeks to really understand that it's not just a new tool or a new workflow. We really do have new AI teammates, and we need to build the team around them.

(6) Cost

we are not paying our AI teammate enough yet. I hope it will change soon..

We are speaking about roadmap in days, new integrations and new products in weeks. Our PM is now a developer on the team…

What's next?

hiring C-3PO to do HR (who can make the intro?)