There’s no microphone on my old laptop (Ubuntu/Firefox) so I tried with my phone (definitely has a microphone), with Chrome browser, and there’s no microphone icon in the top-right corner. 
No worries; just wanted to establish the fact that it seems to need some prerequisites to work in a browser and all of the devices I normally use fail to qualify.
I thought the conversation integration wasn’t enabled by default. Personally, I listed it in my configuration.yaml. Fun little segue lol
would ya look at that, I have it. @123 it’s probably what @bwduncan mentioned. It’s in default config, I have a little scraper that suggests adding integrations to configuration.yaml. Must have added it at some point.
EDIT: It’s not in default config. I must’ve tested it in a beta at some point and forgot about it.
Based on what I have read about the Conversation and Intent Script integrations, I believe they can be used to support a phrase like “turn off the heating at 5am”.
However, if one considers it inconvenient to create an automation then this process wouldn’t be seen as a shortcut.
1 Like
I agree, but setting up the automation once is worthwhile, even if it takes a lot of effort. When I want to go to bed and I just need to toggle a switch while I’m asleep, I need something quick. Sadly I have not had the time even to look at the above suggestions in detail, although they all look promising. Thanks everyone for your support 
Is that a change of heart? Because it’s not the impression I got from your initial posts:
Is the easiest way to do it really to create an automation and then remember to delete it tomorrow? What the heck?
It’s a lot of taps to get from the dashboard to creating an automation
I was hoping for a more “partner friendly” approach.
Maybe slightly! It’s a similar objection to @CentralCommand 's that having to go to the developer tools is inadequate.
There’s also a big difference between having to navigate a long way to edit an automation and having to remember to disable it tomorrow. The latter is unacceptable. I think you have proposed solutions which avoid this problem neatly.
I think what I’m saying is that a “template” which lets me make a one-shot timer automation (like those proposed above) is acceptable, but creating an entire automation from scratch each time is not.
To be clear, none of the suggestions are for a “one-shot timer”. The word “timer” has special meaning in Home Assistant (an entity that counts down a given duration). What you’re after is something that executes at a scheduled time just once (i.e. doesn’t execute on a daily basis at the scheduled time).
You can achieve it with an automation or a script. For added convenience, a blueprint can be used to generate the automation or script. Of course you first have to create the blueprint itself but that’s been done for you (by CentralCommand for scripts and me for automations).
You can also make it even more convenient than using a blueprint by employ Conversation/Intent Script. Then it becomes as easy as just speaking the phrase. However, first you have to create the means of recognizing/acting on the phrase.
Another workaround if you’re in the Google ecosystem and have Google Assistant set up to expose the devices you’re targeting: Ok Google, turn off xyz in 5 minutes
Maybe Alexa can do this too. Don’t have that to verify this though.
How do I import this blueprint?
I created a yaml file, uploaded it to a /config/blueprints/automation
But it doesn’t appear in my blueprints.
I’m doing something horribly wrong here 
Execute: Developer Tools> YAML > Reload Automations
That did it. I wonder why it didn’t do it when I just restarted the HASS. But anyway, it’s working perfectly. Exacty what I needed.