I wrote an internal Alexa skill for [large cola company] a few years back. It allowed them to ask for sales figures based on certain criteria (time frame, category, vendor, ect.)
They never told me explicitly what it was for, but I assumed it was for meetings where it was just easier to ask Alexa than look it up manually. Like the article says, I imagine this will be more popular going forward.
Some of the intents that were part of the skill they requested were just too long to say conformably. I'm guessing it was approved by someone just trying to impress their boss by busting it out during a meeting.
A lot of requests to BI teams are simply “what’s the <metric> for <some set of dimensions> for the last/next <timeframe>”. Literally, thousands of analysts just answer these one liners day in and day out.
More diligent/experienced people will know to ask the “whys”, and recommend better insights, but generally it’s not much more than that and Alex can handle a lot of these provided there is a robust enough dataset/reporting API behind it.
Not convinced, these kind of asks can get mopped up with a web app if there is a robust dataset/API available.
BI teams get used / needed when the data needs fixing and also when the question appears to be of the form above but in fact is of the form "I have a new metric I want to explore across dimensions that are not in the data set"
Wonder if this has been considered: mixing English and “insert your langauage here”.
Also, perhaps not so relevant in 10-20 years, but some folks like my grandma cannot speak mandarin. She can only speak local dialect, which is not covered in any speech recongition package anywhere in the world. The best I can do is make the program to understand a couple simple commands in her dialect, and collect words spoken in her dialect to construct a response.
10 comments
[ 4.6 ms ] story [ 26.3 ms ] threadThey never told me explicitly what it was for, but I assumed it was for meetings where it was just easier to ask Alexa than look it up manually. Like the article says, I imagine this will be more popular going forward.
Some of the intents that were part of the skill they requested were just too long to say conformably. I'm guessing it was approved by someone just trying to impress their boss by busting it out during a meeting.
More diligent/experienced people will know to ask the “whys”, and recommend better insights, but generally it’s not much more than that and Alex can handle a lot of these provided there is a robust enough dataset/reporting API behind it.
BI teams get used / needed when the data needs fixing and also when the question appears to be of the form above but in fact is of the form "I have a new metric I want to explore across dimensions that are not in the data set"
https://h4labs.wordpress.com/2017/12/13/voice-as-a-user-inte...
The Amazon and Google devices should get us to usability and usefulness.
Also, perhaps not so relevant in 10-20 years, but some folks like my grandma cannot speak mandarin. She can only speak local dialect, which is not covered in any speech recongition package anywhere in the world. The best I can do is make the program to understand a couple simple commands in her dialect, and collect words spoken in her dialect to construct a response.
Business : shared space, diverse multiple unknown speakers.