Certain ultimate ends towards which we use HA (“configure mancave for karaoke party”, “winterize home”, etc) involve a combination of automated and manual steps. We used to take for granted that only some of them could be automated, but with the addition of the To-Do list we can have more meaningful higher-level sequences of actions in HA, that combine human and machine-executable steps, guide the human part, and track the overall progress.
This is of a particular importance in a multi-user home or a small business environment, where not every user is technically savvy, and to be really useful some of the sequences must combine human actions or checks (heretofore documented elsewhere if at all) with a series of automatons. This often used to be too fragile or complicated and only doable in full by the one person who’d set up the HA - a bottleneck or, as they now say, a “dumb home”.
Case in point: House Party (guest invites, track confirmations, get supplies, charge outdoor speaker, load mosquito repellent cartridges, ensure propane tank at least half full, ensure roborock vac dock clean water tank full and dirty tank empty, start deep clean, set lights to 100%, retrieve fireworks from storage, retrieve torch lighter from garage, ziptie garbage bags on the patio, set patio lights to 70%, turn on grill, etc etc).
Case in point: Bar End of Service: announce last call, in 5 min lights fade in from service level to 100%, music fade out to 0, close all open tabs, collect all glasses and rack for washing, check bathrooms, count tips, close out day in point of sale system, etc etc.
It’d be good if each manual step could be not just a text item but could optionally include relevant properties such as:
- location of any tools or supplies required to complete the task - to be retrieved from or replaced to
- expected “before” picture i.e. the prerequisite or an actionable / normal “before” view of the thing the current step involves operating
- “action” picture (i.e. a close-up of a Karaoke mic battery compartment or a projector input that are the subject of the action.
- expected “after” picture and/or text defining the condition for what’s required for the valid “done” state of this task
- “on_error” - if anything goes or looks wrong, or prerequisites not met, info and/or automation link.
Regards,
Alex