Chatbots are gaining popularity across the business world, providing a new way for companies to reach out to customers. Since last April, over 11,000 chatbots are available on Facebook Messenger. Companies from Barbie to the Washington Post are looking to capitalize on this new method of communicating. But do chatbots add value to Customer Experience?
The answer to that question depends heavily on your company’s industry. Are you providing weather reports or 401(k) investment advice? Movie ticket times or medical advice? When you step into the world of complex answers and responses, today’s chatbots fall short. Here are a few reasons why:
1. Chatbots today lack expertise
They do not ask questions, explain, or offer advice. They collect information and give formulaic responses. Some of them are basically like a search engine with a voice command. Dialog is based on expertise. For example, when you see a doctor, the doctor asks you the majority of the questions after you tell him or her what’s wrong. That dialog is based on the doctor’s expertise. Automation of true dialog as part of the automation of knowledge work is exactly the same
2. Chatbots still can’t write
Expertise and reasoning are critical, as we mentioned above, but even if a chatbot had the ability to reason, how would it articulate the results of reasoning? How would it write questions? The science of making a computer write like a human being is very complex, in fact, there are only a few companies in this space globally. Automation of smart dialog is possible, where the software actually writes questions based on previous answers. Chatbots cannot do this.
3. Chatbots always provide an answer
I know this may seem obvious, but it’s important. If you ask a chatbot a question, you will always get an answer, but how often is that answer correct? As we have discussed, there are different branches of AI. Chatbots are programmed to always answer. They try to automate 100% of the work and have maybe 70% accuracy.
So if your company can’t afford to be wrong 30%, what’s your option?
A deterministic approach coupled with a smart dialog engine like Natural Language Generation. This type of system can be programmed with best practices, and it is able to say “I don’t know” when it encounters a scenario it is not programmed for. Having this type of control over the system can be critical.
Think about it from a business perspective. Imagine you are working in finance, you cannot have a bot go off the rails like the Microsoft Tay bot, or even less so, giving non-compliant advice. You need to control what the system can and can’t do and that is what a deterministic approach offers.
Incorporating smart dialogue takes the best of chatbots and adds additional value by applying your company’s expertise to scale, adapting the conversation based on previous answers, and knowing when to say “I don’t know”. Being able to quickly answer frequently asked questions, while escalating more complicated requests to the appropriate person, Smart Chatbots has the potential to enhance, not restrict, Customer Experience. This is the future of the chatbot.