First, some data about companies using and considering using chatbots
A recent Capgemini survey of conversational interfaces provided some positive data for companies using and considering using chatbots.
They reported from their 12,000 respondents that:
- 76% of companies have achieved measurable benefits from voice and chat assistants including an increase in Net Promoter Score,
- a 20% increase in first call resolution and
- a reduction in customer wait times of more than 5 minutes
Companies that achieved successful outcomes, however, were those that understood how to best use these tools.
Not properly understanding the advantages and disadvantages of chatbots and their fit to the use case in question
The advantage of chatbots is that they provide a level of convenience that is hard to provide with human support agents both in terms of availability and ability to solve the problem (if the problem is within the domain of the chatbot). Human customer support is mostly a hit and miss process as it comes down to luck whether you are matched with the right person on the first try. For this reason, no one relishes a call to customer support and it is relatively easy to improve upon some aspects of the experience with chatbots.
Not understanding the disadvantages of chatbots can lead to poor outcomes. Again the disadvantages are that unlike humans there are some statements that the chatbot will not understand and not be prepared for. In this case, the chatbot needs to handle the situation in a way that does not further inconvenience the customer.
For example, forcing a customer to repeatedly engage with the chatbot after their initial statement has not been understood is an obvious error. In this case the customer should be given options to either select from a menu of items that may be relevant, talk to a support agent or raise a ticket in the support system.
Another related problem is that the chatbot is inefficient in the way that it communicates. An example of this is voice bots (particularly on the phone) that are slow to get to the point that the user can engage because the programmers found it necessary to make the bot state numerous instructions about use and other pauses and banalities into the bot dialog.
There of course are many best practices involved in building great chatbot experiences that developers should be aware of when designing a chatbot.
One interesting observation of the Capgemini study was that successful chatbot use cases were focused on delivering convenience to the customer. While this is an obvious benefit of chatbots, it was interesting that the customers themselves wanted chatbots to be taken to the next level, in terms of delivering higher levels of personalization, emotional connection, and value.
Personalization and emotional connection, of course, need to be delivered without affecting the convenience of using the chatbot, which is rule number one. Personalization, however, can be powerful.
A chatbot that is aware of the customer history can provide additional convenience and relevance in its interactions that customers will appreciate. In the same way that Uber might know that on Wednesday at 3pm there is a high probability you are going to go to a certain place and therefore place that place at the top of the destination options, a chatbot can similarly use a customers history to deliver additional convenience and relevance in terms of the products or services it offers.
Emotional connection is not just about getting the phrasing right so that the customer enjoys using the bot, but can also be about recognizing the customer and where they are in the customer journey. For example, the bot could reward the frequent customer with a free gift or discount as a thank you for their continued patronage.
In terms of adding value, bots are undersized in terms of providing customer enablement. Personalization coupled with the ability to do things (not just provide information) is a very powerful combination.
For example, saying to a chatbot order me the usual coffee and send it to my office is much more powerful than providing information about where to get the coffee, or allowing the customer to order the coffee but to have to start from scratch each time.
Of course, building sophisticated chatbots requires sophisticated chatbot development tools, such as Botpress, an open-source chatbot development platform.
Companies with a mature understanding of these tools are already greatly improving customer experience with chatbots. There is however a significant opportunity to improve the experience.