Are Chatbots for me?

Published on Jun 30, 2018

Are Chatbots for me?

Making a start and don’t know where to begin? Here are 3 ways to make a step towards a successful chatbot implementation.

Do you have High Inbound Support Volume?

Automation needs to have a large support volume. Chatbots are generally better for ‘in-bound’ support, that is reactive, where the customer to contacts you.

What is high? Once you get to 10 support agents. Add to this 24/7 support; global ambitions and soon the ROI can be very positive.

If your organization caters to consumers (say you are an insurance company or a claims TPA), and you have a 9 to 5 operation with 50 staff. With a Chatbot augmenting the team it is possible to be open 24/7 and to expand. Plus the staff wont have to deal with all the tedious requests allowing them to feel more supported.

Its not all about cutting costs – although in the above example it would be easy to think that you could save upwards of 25% of the effort with a chatbot. Growth and higher levels of support are also achieved. To implement a chatbot it is almost like training a new member of staff – soyou’re your knowledge banks and FAQs and help sheets are all valuable and can be used. Automation is nothing new to you. With a chatbot it will just help you scale up when demand is high – fast!

Are your service requests repetitive?

Training agents is nothing new to organizations, they have training documentation, FAQs, process, guidelines and many tools to help new agents learn and handle the more common customer support requests.

As with many things in this world the 80/20 rules does exist in customer care situations. So, pointing the right answer to the question can be relatively straight forward for 80% of the issues. The remining 20-30% can be harder and can take time or just leave your best people to answer these questions.

If your business has many repetitive requests, a chatbot will help and reduce the amount of support tickets by taking care of requests automatically. This frees up resources for your support agents to deal with the kinds of requests humans are best at.

Can you categorize your support and customer requests?

Broadly there are two types of customer requests: Straight Process or Problem.

  • A Straight Process request is where the customer knows what they want and it’s a core business process for you. The customer might not use the right words as they tell you, but they do explain their wish. For instance, insurance companies receive Straight Process requests like “I want to make a claim” or “I want to get a copy of my policy” or ‘I want to add my daughter to my car insurance’.

  • A Problem request is where customers describe the issue but do not know what the problem is. Examples are requests like “I cannot login to my Policy Account anymore.” Or ‘I can’t read the online system and I can’t change anything’ Or ‘the cleaner broke a vase and won’t pay’ In these cases, the customers’ main want is to get the right solution. On top of this we also have two general classifications of solutions: Self-service vs. Supported.

  • Self-service is a customer can solve requests themselves. A password reset is a simple well known one we all come across. Other examples are Status updates for a process; or balance enquires; or Policy Limits or inclusions. A chatbot can save 100% of human time in these cases. Customers just want the answer and speed and convenience is the key.

  • Supported solution means that potentially an agent is needed to perform some of the customers’ request or task. This can be after the bot has provided the correct solution (e.g. cancelling a policy or paying the actual claim amount). However even in these cases if your bot is connected to your backend systems and can trigger these actions via an API even these processes can be significantly automated.

Using TonkaBI we can integrate your chatbot over time and build the automation deep within your organisation. This is not and should not be viewed as a quick 3 month project. TonkaBI believes that automation is the best type of agile project – identify a well contained need and then solve for that, moving on to the next one – learning as you go. TonkaBI builds all the learning into your organisation adding value as it goes.

At a simple level, chabots are great at solving straightforward transactions and simple queries. They can also help shepherd customers on a relatively linear journey, such as answering delivery questions on an order.

However they can be used for more than just straightforward customer engagement. They have the potential to transform back office processes, freeing-up employees from repetitive tasks to focus on more complex and value-added work. And it can also be used to transform more complex processes, such as claims, complaints and solving for better customer satisfaction.

It is often felt that expensive and cumbersome legacy technologies hold back companies like insurance or broker companies. This can be the case to put off automation. With TonkaBI we have the business and technical expertise to help build an automation layer over existing technology and to update your processes. This can be an effective way to transform a business through automation with lower risk and resource.

At TonkaBI we are well acquainted with many industries and can offer end to end solutions that take advantage of analytics, machine learing and AI. TonkaBI is uniquely positioned to deliver and support globally.

tags #insurance #claims #artificalintelligence #AI # machinelearning

Posted in Insurance AI Blog on Jun 30, 2018