Chatbot-enabled messaging platforms of today can provide very diverse services for users. For instance, Telegram, a popular, open-source messaging platform in Singapore that has integrated payment services within the Telegram Chatbot (https://core.telegram.org/bots/payments). This presents a new opportunity space for e-commerce within Telegram after Shopify (which conquered the web’s e-commerce). With more functionalities and features that can be configured for chatbots in the future, that is a high possibility for chatbots to mimic fully mobile applications, allowing users to perform any services within the chatbot.
In this particular hackathon, I have explored a possibility of using chatbots to retrieve a user’s personal insurance policy document. As insurance policy documents are sensitive, we can use the policy holder’s Know-Your-Customer (KYC) data as Verifiable Credentials (VC) to perform a KYC check before sending over the document to the policy holder.
What it does
This chatbot has basic functionality on providing information on store location(s), and operating hours. The main functionality is verifying and searching for the user’s policy document and transferring the file over. Furthermore, it has natural language understanding capabilities which allows the chatbot to understand the intent of the user when the user is typing in their request messages.
How we built it
We used Affinidi’s API to build the entire VC flow, libraries to develop Telegram chatbot and IBM Watson Assistant to perform intent recognition.
Challenges we ran into
One of the challenges encountered is transferring the response token from the policy holder back to the Telegram chatbot to verify. As the response token that we receive is 11,000 characters long, and that Telegram’s messages only allowed a maximum of 2048 characters long, hence we had to find a walkaround for it due to the short amount of time in this hackathon.
Accomplishment that I’m proud of
Having integrated a number of different technologies (SSI, AI and Chatbot) in such a short time span.
What I have learnt
To learn more about the SSI ecosystem and more about chatbots.
To build more functionality within chatbots to better serve the user. We can also expand the use of SSI into other services in other industries through chatbots.