By Bala Kumar, Jarus Technologies
Long gone are the days when a user would use the manual dial on a car’s navigation system to compose the destination address in a cumbersome manner by selecting one character at a time. Even now, many smartphone users resort to a multi-step process involving enabling the location services, launching the map application, entering the destination, and then selecting the suggested route before even putting the car in drive. Voice-based interactions, however, are quickly becoming a game changer by reducing this minor setup time to just a spoken phrase.
“Navigate to Costco” – just by uttering this natural choice of words to the Google Assistant on a smartphone presents the user with turn-by-turn instructions to the nearest Costco in a matter of seconds.
Such powerful capabilities are made possible by significant advances in natural language understanding and language processing, along with the convergence of a host of other technologies. The ability to interact with computing platforms in this manner, via voice or text modalities, are referred to as conversational interfaces (CUI). CUI is the next disruptive technology that is growing across many verticals.
How is This Interaction Different?
As a society, we have been interacting with computers for a long time via text commands. But those commands follow a very strict syntax that has been pre-defined and requires the user to possess specific skills. Though graphical user interfaces simplified the interaction, often they just act as wrappers built on top of those underlying text commands.
With CUI, users can talk to the system in the most natural form. The system tries to mimic the conversation as if the user is talking to a real human. Further, it is the responsibility of the system to handle some level of ambiguity that is part of natural human communications and to adapt and understand what the user wants to accomplish.
The good news is that the capabilities that are driving voice assistants like Apple Siri, Google Home, Amazon Echo, and Microsoft Cortana, are readily available as building blocks for others to create custom automated intelligent agents targeted to their specific applications.
These custom agents in the form of Chatbots can handle hands-free and screen-free environments that deliver information to the users with speed and ease, ultimately resulting in a friction-free experience.
Why Such a Sudden Interest?
There are several factors contributing to the accelerated growth of Chatbots, including:
- 1. Ubiquity of smartphones.
- 2. Advances in AI.
- 3. Growth of messaging platforms.
- 4. User acceptance.
- 5. Significant reduction in development costs.
For example, Facebook Messenger is used by more than 1.2 billion users worldwide. Other such services like Whatsapp, Line, WeChat, and Kik are growing rapidly as well. Chatbots supporting these types of messaging platforms gain an instant reach and accessibility.
User acceptance of this modality is highlighted by the fact that 20% of all Google search queries on mobile and Android devices are already voice-based. In another study, Gartner predicts that by 2020, customers will manage 85% of their relationships with enterprises without interacting with a human.
Compared to web and mobile applications, Chatbots are essentially server-side applications, and they are relatively easy to develop, deploy and maintain.
Chatbot Building Blocks
Key modules of a typical chatbot application are represented in the diagram below. Users interact with a Chatbot through various messaging applications such as Facebook, Twitter, Skype, and Google. The input voice and text commands are then parsed using the natural language processing (NLP) and natural language understanding (NLU) services.
During this process, the context of the conversation, the intent of the user and other relevant parameters are extracted from the user’s input. Then, the fulfillment service takes the appropriate action to fetch the required information from the backend system and 3rd party services. The Chatbot also manages the dialogue and keeps track of the context for continuity in subsequent conversations.
On certain supported platforms, rich user interfaces involving interactive buttons, images, and menus are weaved nicely between the text interactions. Chatbots can also be designed in a way such that the conversational support can be switched between a human and digital assistant seamlessly.
Chatbots for Insurance and Development Approach
When it comes to Insurance, many self-service features can be easily supported using CUI. Policy summary and billing inquiries are natural choices to be implemented first. Quick quotes, roadside assistance, accident help, first notice of loss, user alerts and notifications are a few of the features that can be considered as a second phase implementation.
Chatbots can be developed in an agile and cost-effective manner. The key steps to implement the Chatbots include:
- Identify the development platform
- Identify a small, well-defined set of tasks to implement first
- Deploy, train, and fine-tune the features over a period of time
- Add additional functionality on an incremental basis
With a simpler interface and easier discovery, CUI offers friction-free transactional capabilities resulting in increased customer satisfaction and retention.
This article was featured in the PAMIC Pulse.
- Josh Constine, “Facebook Messenger hits 1.2 billion monthly users, up from 1B in July,” TechCrunch, April 12, 2017, https://techcrunch.com/2017/04/12/messenger/
- Greg Sterling, “Google says 20 percent of mobile queries are voice searches,” Search Engine Land, May 18, 2016, https://searchengineland.com/google-reveals-20-percent-queries-voice-queries-249917
- “Gartner Customer 360 Summit 2011,” Gartner, March 30, 2011, https://www.gartner.com/imagesrv/summits/docs/na/customer-360/C360_2011_brochure_FINAL.pdf
Mr. Bala Kumar, Co-Founder and VP of Business Development at Jarus Technologies, has over 20 years of experience in developing and commercializing software systems. Jarus specializes in digital and core system transformation solutions for the insurance industry. He received his M.S. degree in Mechanical engineering from the Indian Institute of Science.