This project was an idea from the Sales Manager of one of our existing customer. Previously, he had difficulties on how to keep track which products are sold most, how to keep track the conversation between the sales person and customer, and how to keep track customer’s sentiment of bought products.
The idea is to utilize Natural Language Processing (NLP) and Sentiment Analysis to extract conversation from messenger service, extract products, brands, types, or models information, analyze the conversation’s sentiment based on weighed words scoring. We thus submit this solution for SME SEEDx Development Challenge 2020.
Read more about the challenge here.
BeOne Analyzer Demo Video:
The solution helps the Sales Manager to identify some critical information on his department. By using the web based application, he can extract information from conversation between his team and customer such as how well the sales of each product, brand, type, or model, analysis of each conversation’s sentiment (very negative, negative, neutral, positive, or very positive). Furthermore, for each negative sentiment, he can create follow up by creating service call document on SAP Business One.
Solution Use Case:
The solution is targeted at trading and distribution industries, or sales department at general. The solution consists of:
- A messenger’s conversation extractor to fetch conversations from messenger and convert them into a spreadsheet file.
- A web based application to register words scoring, products, brands, types, and models.
- A web based application as the dashboard to show charts of products, brands, types, or models, and the sentiments of each conversation.
- Follow up module for creating service call document on SAP Business One.
- Unable to track customers’ sentiments toward products, brands, types, or models from conversation between sales team and customers.
- Unable to track which product is being asked or talked often in conversations.
BeOne Analyzer is consisted of some core modules:
- Messenger conversation extractor and spreadsheet converter. Some information extracted: conversation id, subject, date, sender, and receiver. The spreadsheet will then be imported into the web based analyzer.
- A web application to import the spreadsheet file, analyze, and show the data on a single page dashboard. The dashboard will show a table of conversations data, charts for all conversations’ sentiments, and charts showing each product’s proportions of being talked in conversations.
- Menu on the web application to create service call document on SAP Business One for each negative sentiment.
Screenshot of the web application
Screenshot of the web application dashboard
UX Technology Used: HTML5, CSS3
Platform Technology Used: BeOne Cloud Server
Latest Technology Used:
- Natural Language Processing (NLP) for information extraction
- Sentiment Analysis using Word Scoring Algorithm
- Node JS
- SAP HANA
Industry Focus: Trading and Distribution
- Social Media Marketing
- Monthly Workshop
- Beone Channel Partners
Future development will include:
- Automatic fetching and processing so spreadsheet file will no longer be needed.
- More messenger platforms integration as currently the solution is connected to only 1 messenger.
- Mobile application as alternative of the web based analyzer application.
Partner: PT. Beone Optima Solusi
Team Name: BeOne Intelligent Enterprise Team
Team Members: Ronnie Eko Prasetyo, Oki Wijaya, Yoseph Sunarli, Reza Stefano, Brian Billardo
The Challenge is a big chance for us, first participating partner from Indonesia, to begin researching, investing, and exploring the fields of Intelligent Enterprise and Industry 4.0. We are really excited to keep submitting new solutions as we believe a lot of wild ideas about future technology and innovation can be delivered into real things starting right now.
For more information about the development challenge, you may refer to the SME SEEDx Development Challenge 2020.