Chatbots are designed to replicate human dialogue. It’s their USP and their defining feature – they provide a conversational, human-esque customer experience that’s typically more efficient, less expensive and better able to capture and analyse large amounts of customer data than human agents.
However, striking the right tone is more difficult than it first appears. Clumsy and ineffective scripting is often caused by three main factors.
- Too broad an approach to chatbot design
- Inexperience with chatbot scripts
- A lack of focus on your key customer demographics
In the first instance, it’s important to note that there are many different types of chatbot and each is built to fulfil a specific function.
For instance, chatbots can be transactional, information-retrievers, automation specialists, or data collection devices. In each instance, the bot’s script needs to reflect its function.
The second factor relates to specialist expertise and experience. Like any complex technology, chatbots are best designed and built by those who have a good understanding of that technology. While there are platforms that allow almost anyone to build a “functional” chatbot, there’s a big difference between a basic bot and one that’s optimised to fully benefit both customers and your organisation.
Finally, clumsy scripting often results from organisations failing to consider who their audience is and how they communicate. For instance, an airline chatbot that interacts with high-value, business class customers is likely to utilise a different tone of voice to a chatbot that’s designed to help retail customers return an item of clothing.
If your chatbot regularly misunderstands or misinterprets customer intention, it’s going to cause users to reject the service in large numbers. Many chatbot designers implement a fallback message, such as ‘Sorry, I didn’t get that. Could you ask me again?’ to alert the user when a chatbot is unable to understand user inputs.
While this is better than the chatbot pursuing an irrelevant course of action (such as a Local Authority chatbot directing you to a tax advice website when you wanted to report fly-tipping), it’s still frustrating for customers.
This type of fallback message can be improved by getting a little creative.
As well as a fallback message, chatbots can refocus the user on the task at hand. For instance, if a user is attempting to file a fly-tipping report and the chatbot misunderstands an input, it could follow up its fallback message by trying to get the user back on track. ‘Are you ready to submit your report now?’ might be an appropriate response.
Alternatively, the chatbot could be designed to provide the user with a list of tasks that it is capable of handling. This ensures that the customer isn’t frustrated by repeated attempts to communicate with a chatbot that’s never going to understand their intention.
Users will also grow frustrated if organisations rely too heavily on chatbots at the expense of other channels.
While a customer may be happy to enter into an interaction with a chatbot, they also want an easy way out. Failure to provide a means of switching between channels will lock the customer into the chatbot and result in an unsatisfactory experience.
At a minimum, all chatbots should have the ability to transfer customers to a human agent should they request to do so. Similarly, users should be given the option to transfer to another channel should the chatbot repeatedly fail to recognise the customer’s intention.
While chatbots are a powerful automation technology, it’s important to remember that they are just one part of the omnichannel customer service system.
They have their strengths and weaknesses and designers need to play to these – funnelling customers towards those tasks at which the chatbot excels and diverting customers to other channels when it’s unable to help.
Impersonal customer service is a common weakness in chatbots. This is surprising, as chatbots are more than capable of providing a personalised service. In fact, they excel at it.
When we talk about ‘personal customer experience,’ we don’t mean the chatbot greets you by your first name, asks how the kids are getting on at school and whether you enjoyed your dinner party last Friday. We mean an interaction where the chatbot’s behaviour is informed by previous interactions and any data concerning the customer it has at its disposal.
For instance, a retail customer who has recently placed an online order may be provided with options regarding that order. Or a Local Authority customer who is halfway through filling in an online form may be given the option to continue where they left off.
Providing a personal customer experience also prevents customers from having to input the same information over and over again – a major cause of customer frustration across all channels.
Finally, customers typically utilise chatbots because they want a quick and easy solution. If your chatbot is unable to provide a rapid answer to their enquiry, it’s not providing them with what they need.
Chatbots that fail in this regard typically to do so for a variety of reasons, many of which we’ve already mentioned above.
However, it’s also worth considering a wide range of other factors. For instance:
- Is your chatbot targeted at specific users?
- Can it be accessed in the right places?
- Are customers interacting with it in the intended manner?
- Is the interface intuitive and easy to use?
These are all considerations that experienced chatbot specialists will be able to provide advice and guidance on.
There’s been an enormous amount of hype surrounding chatbot technology. As with any ‘next big thing,’ the hype has resulted in a bit of a backlash. The truth is, like any customer service technology, there are well-designed bots and poorly designed bots. You don’t see people pointing out frustrating customer call centre experiences and arguing that the telephone isn’t a powerful customer service tool.
When designed and implemented poorly, chatbots can frustrate users. When designed and implemented by experts, they can revolutionise your customer service provision. Finding and utilising the right chatbot specialist is the best way to ensure that you maximise benefits to your organisation.