Chatbots, powered by artificial intelligence (AI), are becoming a popular and essential application in contact centers by swiftly, accurately, and cost-effectively handling, and helping agents to handle, many customer issues.
But like any particularly rapidly evolving technology chatbots have their issues. To find out more, and how to solve them in order to obtain value from these investments, we recently had a conversation with Christoph Börner, Senior Director, Digital, Cyara.
Q. What are the top issues with chatbots in contact centers?
While businesses increasingly rely on chatbots to save costs and alleviate call center strain, some reports suggest consumers are far less enthusiastic about the shift to automated customer support.
According to new research from Forrester Consulting, chatbots are failing to live up to their potential – leaving 50% of customers feeling frustrated and impacting their overall opinion of the brand.
One of the most common and arguably the most critical areas where chatbots fail is establishing customer intent, which is often due to a lack of sufficient AI training and testing.
Q. What are the causes and the consequences of these problems?
The world is home to 7,100 languages–not to mention countless dialects and all kinds of slang, phrasing, and other styles of writing and speech.
Because of the complexity and diversity of ways in which customer intent can be expressed, even humans sometimes have difficulty understanding it. So, you can imagine how hard it must be to ingrain such skills in chatbots.
For example, a customer wanting to return a pair of shoes could type the request to a chatbot in a number of ways:
- How can I get a refund on my shoes?
- Shoes are too big and want $ back.
- I need a refund on these shoes.
Even though none of these requests use the word “return”, the chatbot must understand that the customer intent is to return an item.
This is where chatbots often stumble, particularly as the need becomes more complex or emotionally-driven. This is why consumers are commonly disappointed by their chatbot experiences. Despite the existence of quite capable technology, the bots don’t always fulfill their intended purpose.
Chatbots typically use natural language understanding, a branch of AI, to interpret typed customer inputs.
But chatbots are only as good as the knowledge that powers them. And due to the complexity and variety of ways that humans can convey their intent, it takes large volumes of data to effectively train a chatbot to recognize the wide range of inputs they must interpret.
Such training data is a set of examples expressing each intent that a chatbot uses to turn into a model for recognizing each intent. By emphasizing target use cases, the training data can teach chatbots to successfully recognize and handle each one.
Q. Please outline the solutions
Human communication is constantly evolving, requiring continuous learning and continuous adapting for the chatbot to assure quality. The only way to accomplish this at scale is to automate the ongoing testing process.
…it takes large volumes of data to effectively train a chatbot to recognize the wide range of inputs they must interpret.
Effective QA throughout a bot’s lifecycle is essential to delivering exceptional customer experiences (CXs). Here are the key testing considerations for organizations seeking to improve the quality and efficiency of their chatbots:
- Create training data from sources such as your call center, or third-party sources.
- Test target use cases, as well as non-target use cases.
- Test the chatbot in the context of the whole omnichannel customer journey.
- Test the chatbot’s understanding.
- Test chatbot escalations to a live chat agent.
- Test chatbot and live chat agent performance under peak load conditions.
- Test the security of the chatbot.
- Monitor the chatbot in production.
Q. What steps can contact centers take to prevent future issues and if they do occur, to resolve faster and more effectively than before?
Companies must train their chatbots how to understand customer intent and continuously test the technology to ensure that, if issues occur, they can catch and resolve them before their customers experience them.
Having an automated and continuous chatbot testing approach, like those outlined in my answer to the last question, improves the quality of the chatbot, thereby ensuring it performs as expected while reducing the efforts of enterprise IT teams.
Organizations should train their chatbots with various phrases in advance so that they can recognize the correct intents.
Effective chatbot testing examines the way a chatbot understands customer intent, and it will do so continuously as the chatbot evolves and consumer input changes. This testing can also be done at scale to ensure the chatbot can perform even at high volumes.
Chatbot testing must evaluate how the bot performs in the various channels it resides, whether that is on the website, in a mobile app, within an IVR, or on Facebook Messenger or other channels.
This comprehensive testing provides value in every phase of the bot development lifecycle, enabling companies to:
- Deliver on their business goals of improving customer satisfaction and reducing costs.
- Mitigate the risk of chatbot fails and negative brand impact.
- Accelerate their chatbot development cycle.
- Increase agent efficiency and improve churn by equipping them to properly direct customer inquiries that can be handled by the chatbot.
Q. Finally, how can contact centers best maximize the benefits of chatbots?
A successful chatbot is defined not only by the technology that powers it, but by a well-thought-out conversational design.
Every chatbot behaves differently, depending on its purpose, topic coverage, and target user. However, there are general best practices to follow when building a chatbot that can improve its quality and lead to a better CX:
- Pre-train your Conversational AI. When a chatbot comes to life, it usually only has a small amount of training data. Organizations should train their chatbots with various phrases in advance so that they can recognize the correct intents.
- Use a fallback strategy. It’s impossible for a chatbot to answer every single question. This is either because the AI is not yet fully trained or the chatbot is being asked questions that it is not designed to answer.
- There are several ways chatbots can handle these fallbacks. For example, a chatbot can capture the contact details of the customer and forward them to a live agent who can then assist the customer.
- Get feedback. Ask users to provide feedback on the chatbot and recommendations on how it can be improved upon. Receiving negative feedback can help identify where there is room for improvement. This can provide a wealth of suggestions and ideas to further improve the chatbot.
- Think about the user experience (UX). When designing a chatbot window, focus on including elements that follow the company’s branding, such as typography or color. Use interactions like buttons, quick replies and cards to give the user predefined options to choose.
- Using such elements can enhance UX because users often don’t know how to write the question or what information the chatbot needs.