Are Chatbots a Preferred Mode of Customer Service? A Point of View | HCL Blogs

Are Chatbots preferred mode of customer service – a Point of view

Are Chatbots preferred mode of customer service – a Point of view
October 16, 2020


The world is slowly switching to digitalization and so are the customers who have chosen to move along with that transformation. The transformation has been exponential in all areas where we perceive interaction between a product/service provider and its customers. area is no exception to this transformation.

Among the financial institutions, insurers were the foremost adopters of technology primarily due to the voluminous data being processed daily. Digitalization had been introduced in the financial institutions in the late 50’s and has evolved to a state where its association is a must to stay competitive.

There are several ways that the customer connects with the other than the personal visit. These are depicted below with their pros and cons:

Point of view

The Pros and Cons of Customer Interaction

Chatbot as a customer-connect channel

Though chatbots are now a common term in the business space, by definition, they are programs based on artificial intelligence (AI), machine learning (ML), and natural language processing (NLP) that enable non-voice conversation as chats through various applications or websites or phones. Apart from the benefits that a customer gets, the implementation of chatbots helps an insurance company in the following ways:

HCL’s ‘DRYiCE Lucy’ learns and adapts to customer needs through smart conversations by leveraging enterprise grade Natural Language Processing (NLP) and Machine Learning (ML)

  • Reduced call volumes in call centers resulting in operational savings
  • Better as no waiting time which will bring in more business
  • Quicker resolution of queries which increases the brand value

Insurers have been facilitating customers to communicate through different digital channels but the usage of chatbots as a digital tool is not finding its due place yet. This content comes as a point of view to highlight the challenges that a chatbot presents; which has become the reason for its low popularity; the potential solutions to overcome them, and also emphasize on the growth prospects in the next five years.

Chatbot challenges and solutions

Insurers across the globe mostly facilitate their customers to contact them through call centers, mails, or IVRS and very few have invested in bots. Chatbot is a digital tool which, if effectively used, brings a lot of benefits to customers and insurers.

Unfortunately, chatbot implementation amongst insurers is not up to the expected level. The global indicators for chatbot usage are-

  • Europe leads with 45% chatbots implemented amongst insurers. However the implemented bots possess limited features across basic, moderate, and advanced level of chatbot capability
  • In North America, though the bot implementation by insurers stands at 28%, the percentage of moderate/advanced bots implemented is significantly high. There is still scope for more implementation of bots capability enhancement.

While we enumerated the chatbot benefits for the insurance company and end customers, it is still a mystery why insurers hesitate to invest in this digital tool. A detailed analysis and surveys taken among the insurers provide the following reasons:

  1. Errors encountered

    Errors encountered Errors encountered

    Challenge:Insurers are concerned about the errors that could be encountered during the course of any conversation which could mar the image of the insurer and thereby lose leads/prospects of a new customer and potentially lose the existing customers as well

    Solution: The bots are programmed in two ways – rule-based bots and AI-based bots. The rule-based bots are programmed with different rules based on different scenarios. But if the user asks a request which is not covered in a scenario, bots will behave in a crazy manner. The understanding and response are stereotyped. The artificial intelligence-based bots are taught to understand the human conversation so that it responds sensibly with the help of machine learning and natural language processing.

    Most of the bots are rule-based as they are built using simple programming and the cost is low. The errors listed above mostly happens in rule-based bots. It is due to the bot picking up only the keywords and responding to them instead of understanding the human dialogue.

    These types of errors are minimized by guiding the user with a standard set of options, following which, the user selects only among those options. But when the user types free text without choosing an option, the bot should be responding politely, asking the user to choose from the options listed or whether they want to connect with the call center representative. This is because no amount of programming can mimic a variety of user interactions. Instead, the bot should be frank enough to admit that it is not human and has got limited options for the user to choose.

    Artificial intelligence-based bots are costlier, but they can learn day by day and be smart enough not to commit the above mentioned mistakes. They will not only be looking for keywords but will also be trying to execute the commands by reading and understanding the sentence as a whole.

  2. Lack of awareness

    Challenge: There seems to be complete lack of awareness for most of the insurers on the benefits a chatbot might bring to them. Compared to other industries, the adoption percentage is lowest amongst insurers across the globe.

    Lack of awareness Lack of awareness

    Solution: There is immense potential for solution providers to make use of this opportunity and build bots specific to certain use cases which will grab the attention of the insurers and bring them quick benefits. For insurance, the use cases can cover the following processes:

    • Billing queries and payments
    • Policy servicing
    • Claims
    • New products enquiry and purchase

    Almost all the customer queries will be in these areas and a few may be on the new products. The mobile version of bots should also be launched as it would be more user-friendly with appropriate security features enabled.

    If the COVID-19 threat exists, there will be challenges for the employees to work in a call center environment. So, the usage of digital channels will be more as the customers will be contacting the insurer for billing enquiry, payment, surrender, loan, and claims. So, a bot with the above use cases can be implemented to gain immediate benefit.

  3. Cost of implementation

    Challenge: Insurers might not be willing to make new investments on digital channels. Although investment in basic chatbot versions will not cost much, insurers might think it as an unwanted expense.

    Solution: This is the best time to implement chatbot as COVID-19 is virulent and cost reduction is the norm of the day. To begin with, a chatbot with a simple use case can be implemented which would reduce a significant number of call center staff and will reduce the customers' personal visits to the insurance office.

    For simple use cases, the cost of implementation will be less than $50,000 which can be recovered within a year. After the initial success, complex use cases can be added at will. Also, simple use cases can be built using rule-based bots instead of going ahead with AI- based bots which is both costlier and complex.

  4. Legacy system

    Challenge:One of the main reasons for lack of implementation is that the insurer is using legacy systems which might not be able to support the digital channels directly.

    Solution:This is a genuine challenge for the insurer but there are several bots which are adaptable to legacy systems. A worthwhile investment on building necessary infrastructure as an integration and middle layer connecting the legacy system would help in achieving the benefits using the bot.

HCL’s offering

HCL’s offering HCL’s offering

HCL’s bot solution is ‘DRYiCE Lucy’. Lucy is an artificial intelligence-powered virtual cognitive assistant, automated to communicate using human voice or chat. Mimicking human interaction, Lucy learns and adapts to customer needs through smart conversations by leveraging enterprise-grade natural language processing and machine learning.

Lucy interacts in the way clients wish to consume the information they need, helping you accelerate time to value, reduce human error, and increase productivity. Lucy has successfully enabled over 600 business and IT use cases, a subset of which is available out of the box and comes along with a cognitive console for enabling powerful integrations.

HCL’s offering HCL’s offering

The business benefit metrics that Lucy has been able to offer are:

  • 30% reduction in service desk costs
  • Improved CSAT of 90%
  • Improved SLA adherence of 30%
  • Improved first call resolution of 30%
  • Rapid implementation using out-of-the-box integration adapters

Future Potential of Chatbots

The growth projected for chatbot implementation is 29.7% by the year 2024. With self-service option as the preferred mode of customer service connection, it will not be a surprise if chatbots become a common and basic service entry point for financial services.

Also, studies indicate that increased activity expected in the chatbot space in the APAC region compared to other geographies as the region is yet to realize the potential of chatbots across customer services.