Qustomate is a Customer Relationship Management (CRM) application developed by the C23-PR507 team and supported by Bangkit Academy led by Google, Tokopedia, Gojek, & Traveloka
© C23-PR507 Bangkit Capstone Team
| Name | Student ID | Path | Contact |
|---|---|---|---|
| Muhammad David | M313DKX4618 | Machine Learning |   |
| Muhammad Rey Renoult | M312DKX4688 | Machine Learning | |
| Moh. Bayu Aji Prasetyo | M038DSX1879 | Machine Learning | |
| Ananda Satria Ariyanto | C172DSX1973 | Cloud Computing | |
| Isa Tarmana Mustopa | C172DSX2370 | Cloud Computing | |
| Muh. Dhiyaul Haq | A172DSX2077 | Android Development | |
Intense competition in the product and service industry makes companies not only rely on product quality but maximize service for customer satisfaction. According to the 2020 Sea Insights Survey, 54% of MSMEs are increasingly adaptive in using social media for sales. One of the business strategies for managing customer relationships with companies is integrated Customer Relationship Management with Omnichannel. According to the Aberdeen Group, companies that use Omnichannel CRM can retain 89% of their customers compared to those that do not use only 33% of their customers. Because CRM can streamline and speed up customer handling. However, CRM is currently not widely available on the Mobile App platform. Therefore, we developed an Omnichannel CRM application called Qustomate to increase sales and customer potential for MSMEs which is supported by chatbots and sentiment analysis.
We built this application because we saw the potential for MSMEs in the Indonesian economy to contribute 61.07% of GDP. However, in this case, MSMEs do not yet have access to broadly develop their business. Some MSMEs still use the old methods of manual recording, replying to communications via telephone, SMS. The business model is inefficient, causing business development to falter. Based on this, our team developed a platform called Qustomate to increase the potential of MSMEs in Indonesia.
Machine Learning Documentation
Build a machine learning model to determine the polarity of customer chat using sentiment analysis with the LSTM algorithm. We preprocess the dataset of customer using tokenizers and word embedding with tools like Tensorflow, NLTK, and Sastrawi as our base model will be in Indonesian.
Using Firebase services such as Firestore Auth to create Login & Register, also using Firestore Storage as our app database. We also create the required API endpoints using the Python/Flask language, APIs such as DashboardAPI, authAPI, managementAPI, and messageAPI, to be connected to Mobile Development later, while other APIs such as sentimentResultAPI we use to process models that have been made by Machine Learning. For the deployment of these APIs we use App Engine.
Mobile Development Documentation
Create applications for MSME owners and their employees (sales). making applications is focused on employee use first. in the application there is a dashboard page to view sales performance (logged in users), incoming chat columns, client management, and also settings. in the dashboard there is a sentiment analysis chart to display positive or negative customer responses.
| Bangkit Academy led by Google, Tokopedia, Gojek, & Traveloka | Gunadarma University | Republic of Indonesia Defense University | Sepuluh Nopember Institute of Technology |
 |
 |
 |
 |
- Markus Andreas (Tech - Machine Learning/Data/AI, Tech - Android/Mobile Development, Project Management, Ideation/UI & UX Mentor)
- Muhammad Raihan Nismara (Tech - Cloud and Back End, Project Management, Ideation/UI & UX Mentor)
- Iqbal Ahmad Dahlan (Republic of Indonesia Defense University Advisor)
- Faturrahman Irwansa (CC-33 Mentor)
- Steven Adi Santoso (ML-52 Mentor)
- Mohammad Tauchid (CC-50 Mentor)