Best strategy to collect measurements


I would like to create an assistant, wich collects my spoken measurements and put it into an google sheet.

This measurements have their “position” name e.g. AB; AC; AD… and so on.

My goal is to collect data when user said “AB 57” or “AC 75”.

I am not looking to have a two stage input like measurement name → measurement.

I tried this with an intent “measurement” which consists of only two entities e.g. {measurement_name} {measurement_value}. but is not working very well, because voiceflow as also google interpretate this short measurement names as something diffrent. It is difficult to get this short names, especially some name are real words in my language (e.g. ab).

How can I improve this situation, that the name and value of an measurement is instantly caputered?

Best regards

Hi there, the more synonyms and utterances you’re able to define and structure to add to your NLU/Interaction model the more accurate it can get to your parameters. Let me know if you’re still facing issues with a well-defined model.

Hi there,

that is maybe the issue. The Utterances for a measurement is very limited. You may call it ab, a to b, from a to b. The result it seems to me always the same. Due to short name there is a high chance to misinterpretate the call and you get stucked in some loops.

Also the goal is to get the measurements in short phrases into the system. To minimalize the time efforts with recording the data.

I thought maybe there is an other option. :confused:

Have you tried using spelled out versions of your target words?

So in Alexa’s American English accent, Ay, Bee, See, Dee for A B C D, or Ay Bee for A B.

FYI: I test how words sound (and may be heard by Alexa’s NLP engine) using either VF’s text to speech button on the speech block, or Amazon Polly via AWS when I have to do a lot of testing like this.