4 min read
No matter the number of years, one debate still persists in the insurance industry: that of generalist vs specialist software solutions for customer service operation (including support for claims, renewal, and sales ).
Specifically, in the realm of chatbots for insurance, the inevitable question always arises sooner or later: “Which of the two is the best option?”
Is the “octopus” conversational AI platform which is multi-channel or the “honey-bee” one that focuses on executing critical insurance processes? In this blog, we explore the answer to this question.
The crux of the matter: Use case is king
The short answer to the above question is that there are no absolutes. There are multiple solutions in the industry: some are better suited to address a certain type of business needs and yet others are fitted to address completely different ones.
Within the domain of chatbots, we have two categories:
- Generalist chatbots which aim at replicating an FAQ section but cater to no specific process
- Specialist chatbots, where the conversation is created around particular processes
Therefore, the very first question when choosing between the two should be: what do I need the chatbot to do? Or in other words: what is the use case?
In the context of insurance, customer service chatbots built on conversational AI platforms can be considered generalist solutions whereas claims FNOL, renewal, and quote & buy ones can be categorised as specialist solutions. Let’s look a bit more deeply at both options.
Generalist chatbots
Normally, generalist chatbots have a wide spectrum and short interactions. They might include voice interaction such as IVR too.
They have a wide spectrum because their purpose is to focus on and cover multiple queries. For example, in the case of insurance customer service, every customer is coming to a website with a particular query and phrasing it in a unique way. The chatbot or assistant will often start with “Describe me what you are looking for” and then follow with “For payment enquiries, please wait for an operator to be available.”
The interactions are short in nature since most of the time, these requests are Q&A types and stem from customers’ knowledge gaps about policy types or terms.
To manage such queries a powerful system is needed. This is where advanced technology like Natural Language Processing (NLP) comes into play. Conversational-AI, which is heavily relying upon the NLP paradigm, is specifically designed to capture the customer’s intent and suggest the next immediate action rather than solving the process end-to-end. Owing to their usability, 77% of executives have already implemented and 60% plan to implement conversational bots for customer service and sales as per Accenture.
Specialist chatbots
On the other hand, specialist chatbots cover a specific process in-depth, which translates into a narrow spectrum and longer interactions.
A prime example here is the claims FNOL processes, where a policyholder needs to share all the details of the accident: there are a number of specific questions he needs to answer in order to successfully report it.
This is where Conversational Process Automation (CPA) comes in handy. CPA is designed to execute a specific business process in depth. Not only can CPA employ chatbots to capture the necessary information without waiting time for all parties, but it also implements other automation strategies to enhance back-end processes. Following the FNOL example, CPA can triage claims according to data captured, or sometimes even payout low-value and fast-track claims if so designed.
So, Generalist or Specialist?
We’ve seen how generalist chatbots stretch horizontally as opposed to specialist chatbots that develop vertically.
It’s the same difference between a generalist and a specialist doctor or a receptionist and a technician. The receptionist doesn’t have the ability to directly deal with your request, but it might point you to the relevant person (a specialist) that can help. Akin to the health sector, individuals often see a generalist or general practitioner in order to be transferred to a specialist doctor at a later date.
As pointed above, the answer comes back to determining what the use case is. When it comes to insurance, we’ve found that individuals and policyholders have specific and recurring needs that are best addressed end-to-end from their first digital interaction. It gives immediacy and superior customer experience to the individuals while ensuring that the process is executed as per requirements for the insurance company.
Irrespective of the choice, there are two universal rules that both chatbots need to comply with:
- Never underdeliver. Never delegate the chatbot with tasks that it is not able to execute or not execute well. Be realistic about the technical limitations rather than frustrating your customers by demanding more from the chatbot than it can deliver.
- Never overpromise. In order to manage customers’ expectations, always be transparent about the chatbot capabilities. Failing to meet customer expectations is two times more destructive than exceeding them as per KPMG. Thus, if you cannot always provide a stellar experience, make sure to provide one that is expected and does not overpromise.
When these paradigms are ignored, the chatbot solution ends up looking like the hilarious Catherine Tate Show’s information woman, at your brand’s expense.
To wrap it up
Following Vanessa Redgrave’s thought:
Ask the right questions if you're to find the right answers
It stands to reason that the best starting point is to ask yourself and determine what your needs are. Following the trail of the answer, define the use case more clearly and then choose the most appropriate technology to handle it.
Since there will be significant investment in this technology - both in terms of time and money - it is advisable to take your time with the process. But, once you’re sure what your business and vision need, don’t hesitate to allow technology to take you there.
