The client is a leading lending market platform that connects borrowers to lenders for mortgage loans and refinance offers. Through their web-based platform, the portal collects loan requirement details from the borrower and finds relevant loan offers. The client wanted to explore chatbot-based platforms to reach out to today’s wider audience of borrowers who prefer instant access to loan offer services on their mobile devices. So, the client developed chatbot support for Apple Siri. However, to cater to the growing number of Android users, the client wanted to provide chatbot support on Google Assistant, accessible on mobile phones, as well as Google Home devices. To implement this upgrade, Xoriant experts were brought in based on their technology expertise in creating personalized digital experiences with AI and NLP-enabled voice assistants for Android and iOS devices.
The key objectives included:
- Allow borrowers to interact with the chatbot on mobile phones, as well as Google Home devices, through Google Assistant.
- Provide users with best offers from multiple lenders, based on required parameters such as zip code, down payment, loan amount, credit ratings, and so on.
- Enable borrowers to interact in textual as well as voice formats with the chatbot.
- Connect chatbot to various in-house APIs to fetch offers, submit leads, etc. to appropriate parties.
- Configure chatbot to provide inputs to support strategic decision-making and analysis of customer preferences and behaviors.
Xoriant Solution | Key Contributions
Based on discovery sessions with the client, Xoriant experts documented the desired behavior and functionality of the chatbot and implemented it to be easily accessed by borrowers who are Google Assistant users.
- Developed the chatbot using Google Dialogflow and the Actions on Google integration.
- Created a pre-defined conversation flow that allowed the bot to drive each conversation with the user(borrower).
- Enabled speech and textual interactions.
- Designed and implemented horizontally scalable architecture to handle seasonal variations in traffic.
- Added a parameter correction module to overcome limitations of Google speech-to-text conversion and Dialogflow to avoid incorrect voice interpretations.
- Integrated with internal loan explorer platform to fetch the best deals, generate leads and allow the user to connect to the lender.
- Enabled the chatbot to maintain user context during the conversation to make it more effective.
- Implemented relevant logging to enable custom analytics and business intelligence.
- Tested the chatbot on various mobile and Google Home devices for recognition of certain accents, utterances and sentence formats, including American English.
High Level Architecture
- Enabled lenders to capture more leads from an expanded user base of borrowers who prefer Google Assistant voice-based interaction for service requests instead of installing apps or visiting web pages.
- Delivered an innovative chatbot user experience for borrowers with a well-designed conversation flow for accurate response to voice commands.
- Allowed borrowers to use voice-enabled chat to share details for loan inquiry and initiate requests for loan offers via email or a call back from a live support agent.
- Empowered the client with an intelligent, custom AI-based chatbot to gather insights for analysis and decision-making for region-specific service requests; specific loan types; chatbot request and response data, etc.
NodeJS | Dialogflow | Actions on Google | MongoDB | Redis