Published: Nov 19, 2020 by Jenno Vink
The story behind the Gary Vaynerchuk search engine and integrating it with google assistant
A few weeks ago I saw this video of Gary Veynerchuck:
I was a bit disappointed at the moment he said:
GaryVee Bot doesn’t exist yet…
In another (non published) hobby project, I created a google assistant skill which picks a random item from my todo list (weekmenu) and reads it to the used. Me and my girlfriends always like to plan all the evening meals on sunday for the whole week. Letting a smart home assistant decide what to eat, based on a list was quite fun to make.
Based on that I already had some experience creating simple google assistant skills, I saw this as a fun project to do.
Quick demo
Techniques used
For this Proof-Of-Concept, I gained experience using:
- a micro-framework called Flask
- ngrok for exposing my local flask app to the outer world.
- fetching results from API calls using python
My POC works by simply extracting everything an user said after ‘ask gary vee’. So in case of the video, this was ‘how to get success in life?’. This value is used to make a call to this endpoint:
https://search.garyvaynerchuk.com/api/search?page=1&search_value=how%20to%20get%20success%20in%20life
Then the API results are fetched, and formatted into json:
return {"fulfillmentMessages": [
{
"card": {
"title": responseTitle,
"subtitle": responseSubtitle,
"imageUri": imageUrl,
"buttons": [
{
"text": "click here to watch",
"postback": youtubeUrl
}
]
}
}
And there it is. I think that for a POC, it works pretty well. The next steps would be hosting this python script in a cloud environment such as Azure, adding error handling and publishing it to google