8 second response time is not sufficient for IoT use cases
I'm working on an IoT Industrial use case for which I need an "open" response window to communicate with the Alexa service. For a successful proof-of-concept, a user needs to remain focused on the task at hand, while Alexa responds to voice cues in the background to drive IoT devices. The 8-second user response limit is unusable in this scenario.
Currently, extending the the session timeout duration is not a supported feature of Alexa Skills Kit.
Yet, I hope that a skill is closed only when I give certain command, so that I don't need to activate it again. (I only activate one skill at a time)
Jannik K. commented
When it comes to chess against a good cpu player, often it can take more than 8 seconds to respond for a cpu player.
Thats why it is currently not possible to develop similar games.
It would be cool to have Alexa poll for more than 8 seconds until the game can actually respond properly.
Josh Broadhurst commented
Josh Broadhurst commented
This could also be useful for waiting on AmazonPay directive processing if using a mobile Alexa device like the Echo Auto. The skill I am working on is targeted to drivers, but if network connection is not ideal (driving through tunnels, driving in congested city areas), the processing might take too long and the session will end before payment has finished processing (I have seen this during some tests). Eight seconds is not long enough to be reliable.
Chris Drew commented
Currently (10.2018) there is an 8 second timeout for Custom Skills that is standard across all interactions and isn't customizable.
We are aiming to develop a skill that would allow users to opt-in to a skill that is not constrained by time. I.e. the skill will be designed to listen for an unlimited amount of time.
The application for an unlimited amount of time is educational: for families with children with learning disabilities or language development issues there is an opportunity for specialists to be able to measure and analyze a family's "language environment" by using Alexa. Specialists would be able to access the data to proscribe intervention plans based on objectively collected data (as opposed to human recorded data or intermittently collected recordings/transcriptions).
There would be strict permissions and privacy controls and very clear opt-in language for this skill. It would likely even be advised that Amazon provide restricted access to this skill based on the intended application for each developer's use case.