How to review low-energy windows with Hydrogen
Low-energy windows are one of the fastest ways to turn raw tracking into useful interpretation, especially when they repeat across normal workdays.
- Repeated windows matter more than one-off bad hours.
- Hydrogen gets stronger when you ask about recurrence, not isolated dips.
- The right follow-up question is usually about what changed around the window.
A bad afternoon is not a pattern
A single low-energy afternoon might just be noise. A repeating low-energy window across comparable workdays is something else entirely. One is anecdote. The other is a problem with a shape.
The trap is to ask Hydrogen *"why was Tuesday afternoon bad?"* and treat the answer like a verdict. The better first question is whether Tuesday-afternoon-bad is something that has happened before. Recurrence first; explanation second.
Three prompts that work
The prompts that work best for low-energy windows are narrow enough that Hydrogen has something concrete to look at, and broad enough that it can come back with something you did not already know.
- Which low-energy window repeats most on workdays?
- Compare my last 10 workdays and show where energy dips cluster.
- What changed before my lowest-energy window this week versus last week?
What to ask once you find one
Once a repeating window is on the table, the next step is to ask what changed around it. Total workload, activity decline, irregular tracking, missing wearable context — the candidates are limited. Hydrogen is good at this when you point it at the window directly.
The follow-up question that works almost every time is some version of *"for this window, compare X across recent weeks."* You are giving the agent the where and the what; it just has to do the comparing.
