Company Updates

Akvelon’s New Social Chatbot Uses AI to Increase Employee Morale and Communication

blank

In an effort to increase employee engagement and involvement, a team of our developers created an artificial intelligence powered “Social Chatbot” that learns our employees’ interests and preferences, and sends them relevant company news and social media content. This Social Chatbot was designed to ensure that our network of employees felt connected to our company and to their peers through the personalized content they all receive.

blank

Here is how the Social Chatbot works

 

First, download the chatbot here and follow the instructions to add the AkvelonChatBot account to your Skype contacts.

blank

 

 

Once you have added the bot on your Skype account, open a chat with the bot and answers its “get to know you” questions.

blank

 

 

Once you complete this process, the bot will create an account for you and give you the option to add more information about yourself.

blank

 

From there, you simply ask the chatbot to show you news based on your interests (you can even specify where you want the news to come from). There is even an option to view each post individually and an option to repost the news to your own social media.

blank

 

You can also ask the bot to show you specific social media posts from the company, again giving you the option to view the individual post and share it on your own social media page.

blank

How We Made the Bot

With the help of Microsoft Bot Framework, our chatbot is able to use AI to interact with our employees and learn about their interests and preferences. Our developers also created AI that analyzes each chat and filters out any news or posts that the user has already seen. The text similarity model was developed using a hybrid Corpus and Knowledge-Based similarity method that utilizes a cosine similarity measure. Explicit Semantic Analysis and NLTK library were used to implement the model and the WordNet lexical database was used to gather news from different news sources and social media accounts. A group of linguistic two-layer NN models are used to look up specific domain related keywords.

blank

 

Results

We have already noticed an increase in employee interaction and we expect involvement to continue to increase the more our employees use the bot. After the success the bot has had at our own offices, we are currently developing a chatbot system for another company as well.
The chatbot is constantly being improved and right now our developers are refining the text similarity algorithm using a recurrent neural network as well as teaching the Chatbot’s AI to prepare summaries of news and social media posts based on the user’s preferences and previous behavior.

We are always pushing ourselves to create and work with top notch technology and while we pride ourselves in the work we produce, we know that everything we have is because of our employees. The creation of the Social Chatbot was one way for us to make sure our employees felt connected not only to the company but to each other.