split Google Assistant output

Question

So I basically work on a voice assistant for a lecture at my university. It should hold the lecture. A user can listen to the lectures. He can also listen to a specific slide or subject. I am right now using Actions on Google with Dialogflow. My webhook then should return the right content to be outputted. How I will exactly do this, I am not sure yet. But I first tried to output the whole lecture text in case the user wants to listen to the whole lecture. But the google assistant will not speak and the response you can see is not complete. At the end of the text you see "...". I guess it is because it is too long. For Alexa I know that the outputSpeech response cannot exceed 8000 characters. I did not found how many it is for Google (Also I did not get an error like in Alexa?).

So I think I have to split and ask the user after each slide if he wants to continue. But this could be annoying. Is there no way to trigger my webhook again after the response without userinput (like yes)?

Also it would be nice to skip back and forth. Same as when google assistant reading audio books: https://support.google.com/googlenest/answer/7573974?co=GENIE.Platform%3DDesktop&hl=en

But we cant realize such things or?

Thanks in advance for any suggestions. Note that it is not restricted to Alexa or Google assistant. I can do anything as long as it is a voice chatbot.

Answer 1

The platform is designed to be conversational, so long blocks of text may not work well. There are some potential ways to make your Action work a lot better in the format.

• If you can break up your text into smaller chunks, it will be easier to digest it. This has the downside of requiring regular conversational intervals of the user responding with some sort of acknowledgement to continue. This does give them the opportunity to ask to repeat.

• You can convert the text using speech synthesis into audio and use a media response . This has the benefits of letting you save the user's point in time, so they can resume to that section in a later session. However, a media response will not be accompanied by text.

You may want to consider adjusting your design to fit one of these conversational models.