Great question – honestly it’s something I’m struggling with right now.
Straight up with you: Currently the AI silently re-prioritises in the background based on priority score. No clear conflict has come to light yet.
But I remain stuck on this design choice. For reactive-heavy work – which is really your due diligence sprint scenario – silent reprioritization is convenient, but the moment the user realizes the deadline has passed without them even knowing, trust is gone. That’s a very high price to pay for zero-setup.
So the direction I’m leaning towards is to divide conflicts into soft vs hard:
- Tender (Can be solved by changing time slots, clarifying priority differences) → AI handles silently, no notifications
-
difficult (Can’t really fit, deadline conflicts) → Always comes out clearly, let the user decide
The principle I want to emphasize is: Never silently miss a deadline. Everything below that line becomes automatic.
In parallel, I’m moving towards per-user duration estimation – learning how long a given task type actually takes for that person. If we can accurately estimate the duration of a task, conflict detection automatically becomes faster. The beta is partly about gathering that data.
Interesting to know how you drew the line between “taking away the decision” and “letting the user choose” on Dishrolls. Meal plans can be “generate completely automatically from this week’s ingredients” or “surface 2 options, let them choose” and they feel like very different products. Would love to know what you thought about that deal.
<a href