Inspiration
- Our first thought was to have a functional and useful application that will be used on a larger scale by our colleagues.
- Since we have grown a lot in the past period and we are becoming more and more mature, we are adding more and more processes or good practices. It is not easy to know where they are and always know how to find them.
- The most used tool in Zitec is Google Chat. This is why we wanted to have a chatbot that resolves the above problem.
- this idea is also based on our personal experiences and needs;
What it does
- It is a chatbot found only inside our organization that answers questions related to our processes, and good practices, or gives information on where to access different resources.
- It also has good jokes inside!
How we built it
- With a lot of coffee!
- Google Cloud Functions that were written in Node.js 18.
- The application reads from a pre-configured spreadsheet. The spreadsheet contains the most common questions and answers regarding our internal processes and needs.
- The application returns results when it has a minimal level of trust of 40%.
- Given the fact that it reads from a Google spreadsheet, it means it can be easily maintained by our team, and the source can increase with ease.
Challenges we ran into
- Making the bot read from a Google spreadsheet.
- Having an accurate match between the user's question and the appropriate answer.
- We didn't manage to finalize the Google Chrome extension for the bot due to extension limitations.
- Higher learning curve on the new technologies we ran into. It was a first for the entire team. We chose to work with something unfamiliar.
Accomplishments that we're proud of
- Our application works!
- We collaborated very well together, and we had a lot of fun.
- We have an awesome new personality on the team - the one and only "ZAssistant" - that has a nice personality and an awesome logo!
- A set of 80 answers that contains most of our FAQs.
- We also build the front-end for google chrome extension, however we didn't manage to have the back-end for it fully functional.
What we learned
- Google Cloud Functions.
- Google Cloud API.
- IAM permissions.
- How to build a chatbot.
- Some React.
- More Node.js.
- New stuffs about Zitec since we had to gather all the data.
What's next for ZAssistant
- Finish the Chrome extension.
- Make the bot send messages from time to time regarding good practices or advice.
- Integrate the bot with the Google Cloud Search function and give it access to our internal data sources (Google Docs, Drive, Sites, etc.)
How to test the application?
- Go to Google Chat.
- Search for the ZAssistant App. It's a chatbot available inside our organization.
- You can interact with it by two ways: 3.1. Directly ask it for something related to our internal resources. 3.2. Add it to a room and ask it there with "@ZAssistant App QUESTION". !!Disclaimer - our set of questions/answers are in RO
Sample questions:
- care este parola de la wi-fi?
- cum adaug un key result?
- cum cer o adverinta?
- ce framework-uri de testare avem in zitec?
- ce departamente avem in zitec?
- cum pot sa particip la 1024?
- on the main branch we have the code used for the google chat
- on the front-end branch we have the code used for the extensions, however it is only the front-end, without any back-end https://docs.google.com/presentation/d/1w8KUXUlN22pdO1oDMook8rhdF56YnUBk1RjokkFCAGk/edit#slide=id.p
Built With
- amazon-dynamodb
- amazon-web-services
- aws-lambda
- figma
- google-cloud
- google-cloud-functions
- jquery
- node.js
- react
Log in or sign up for Devpost to join the conversation.