Hi Pierre-Gilles,
Already: congratulations 
Great news, enjoy these first months, they are magical.
On the roadmap, I find the order just right and I wouldn’t change anything. Priority 1 is the one with the best leverage effect: as long as an external integration cannot display its own widgets or expose scene actions and triggers, we are stuck.
I started a sort of alarm revamp here, but that can wait for your return in 2027.
And on the 2027 goal: being able to live off Gladys is also what secures the project in the long term for all of us. I will switch from HA to Gladys by the end of October (the comeback) and I will be very happy to support this project again 
Cyril