Basic chatbots are easy to build. As in, the heavy lifting of creating a Natural Language Processing (NLP) system has already been done by tech giants. These NLP plugins are owned and maintained by the likes of Amazon, Google, Facebook, and Microsoft. However, to create a chatbot that is specialised in handling customer onboarding and financial transactions, one would require expert domain knowledge. Therefore, it is wise to pick a vendor with significant number of years developing financial technology solutions.
While evaluating a chatbot vendor, it is also important to bear in mind that they are probably using something like Dialogflow, an NLP that is powered by Google’s machine learning and built on Google’s infrastructure, or Wit.ai, an out-of-the-box NLP plug-in for developers, owned by Facebook. Very few chatbot vendors in Malaysia will be showcasing their own proprietary NLP.
Meeting Informational and Transactional Needs, Today
According to Gartner’s Market Guide for Conversational Platform (published 18th June 2018) while assessing chatbot vendors, one should focus on solving current and immediate needs rather than predictions of future needs. Because the chatbot vendor landscape is poised to change drastically over the next two years, and Gartner predicts most companies will need to change their chatbot vendor within this timeline, and therefore it is prudent to have an exit strategy. Because of this, one should take care to ensure that strategic assets such as training data used for intent recognition, and the integrations required to carry out transaction logic, are transferable to a new vendor with minimal effort. Thankfully, Anny fulfils this requirement by having solid integration capabilities, and straightforward access and maintenance for strategic assets, making them easily transferrable to a new vendor in the future.
So, what are the immediate short-term problems in the financial services industry which can be solved by deploying a chatbot?
In banking and finance, a chatbot’s role in enhancing user experience can be broadly divided into two categories: informational and transactional. A transactional chatbot is easier to build and implement because conversations about financial transactions are usually more structured.
Also, it is important to note that a chatbot is designed to handle initial interactions and then ultimately direct the consumer to your product. So, it is vital to have an actual product that the chatbot is linked to.
Anny, Juris’ proprietary chatbot, is specifically engineered to cater to financial service-related questions and transactional requests. Anny comes with strong integration capabilities with Juris Access, our digital customer-onboarding platform. Anny can also be integrated to your existing solutions or as a stand-alone bot that answers queries relevant to a specific product or service.
Juris Anny Core Functionalities
We have combined state-of-the-art NLP technology and our 25 years’ worth of experience catering to financial institutions to create a bot that makes sense for banks and credit leasing companies. Here’s a quick overview of Anny’s features:
Natural Language Processing (NLP)
- Detects English and Bahasa Melayu (BM) languages.
- Understands a mix of both languages.
- Understands short forms of words from both languages.
- Follows user’s written language as response.
- Has a response accuracy rate of 80%.
- Understands context, entities, and keywords used by the user.
- Learns on its own in real-time based on chat history, patterns, and new data.
- Identifies wrongly tagged data and categorises them.
- Grows word library with minimal human intervention.
- Has an intelligent response to uncertain user intent.
- Supports live chat platforms (INteractive INtelligence [ININ] version 2015 R4).
- Supports social media platforms (Lithium October 2018 release, Salesforce
Social Studio version 44.0).
- Analyses the users’ sentiments and trends based on the chat history.
- Categorises questions based on chat context.
- Provides analytical solutions for improvements in user experience.
- Supports XML and JSON formats.
- Supports RESTful and SOAP protocols.
We also provide end-to-end management of the customer from loan origination to collection. For more information, you may refer to: