An AI chatbot can be a solution to many e-commerce problems. But they are not a panacea, and they cannot function without the proper support. For example, according to a recent Forbes Technology Council article, “Not having enough data can result in a poorly built AI system.”
E-commerce retailers must also understand whether or not chatbots are truly solutions to the problems they are facing. For example, a chatbot that gives generic responses to simple questions might not address the problem of agents taking too much time trying to sort out more complex issues for customers. That issue might require a more personnel-based solution, such as improving customer care representatives’ problem-solving skills or providing them with the right databases to access information.
For these reasons and others, e-commerce chatbots were not always viewed as a success when they first appeared. In the following sections, we explore why chatbots failed in the past and sometimes still do, the implications for e-commerce operations, how they can do better, and the future of chatbots. But first we examine why all this analysis matters—the benefits of e-commerce chatbots.
Customer Benefits of E-Commerce Chatbots
Chatbots can streamline customer support by serving as a first-line response generator for commonly asked questions. For example, an e-commerce site might set up a chatbot to respond to questions like “When will my purchase be delivered?” or “How do I return an item?” Customers can get these, and similar questions answered quickly and accurately with no wait time, as would be the case if they were trying to speak with a human representative.
E-commerce chatbots used for this purpose can be deployed across multiple platforms, including the company’s website, social media channels, and other communication networks. The information provided by the chatbot is the same across these outlets, giving customers a sense of consistency, which contributes to a positive customer experience (CX). Further, chatbots require no waiting time. Customers can access them 24/7/365 and get an immediate response. Moreover, chatbots can often respond in multiple languages.
As chatbots gather information from customers in the conversation process, that information can be retained and used for future interactions, or used immediately if the customer needs to speak with a human representative. This feature makes customer care interactions more seamless, further contributing to a positive CX. Chatbots can also be used in the purchasing process, making shopping more personalized and interactive.
Another use for chatbots is as shopping assistants. In this role, they can do things like personalize product selections, upsell on related products, answer questions, remember shipping and payment information, and easily handle post-purchase inquiries. These actions provide even more opportunities for e-commerce operations to deliver high customer satisfaction.
Business Benefits of E-Commerce Chatbots
Chatbots can be used for a wide variety of actions that help companies save money and increase revenue. For example, chatbots can be set up to remind customers about items that have been placed in a cart but not purchased. They can help e-commerce operations reduce the number of returns by offering exchanges as alternatives and gathering insights about why customers are returning products to fix those issues and avoid future returns.
At a more concrete level, customers that appreciate such personal customer service touches are likely to spend more. According to various sources, a large percentage of businesses have indicated that personalization has increased their conversion rates and led to an increase in revenue.
When customers are happy, companies benefit. According to a recent article appearing in Financial Express, “Streamlining customer support and increasing customer retention are two of the best ways for a business to grow.” This advice is even more true as customers are coming more and more to expect this level of service.
Examples of Successful Chatbot Implementations
A recent Sinch blog post explains that, taking advantage of the business benefits of chatbots, many companies are using them to grow their business. They include pizza chain Domino’s, which, according to the post, “has one of the most sophisticated chatbot experiences.” The chatbot is available on the company website and also on Facebook Messenger, Amazon Alexa, Google Home, Slack, and Twitter. This range provides access to customers across a variety of channels they might be inclined to use.
The Domino’s chatbot helps users through the ordering process, recalls previous orders, and handles order payments. The post notes, “With this seamless chatbot experience, Domino’s keeps customers coming back.”
Another example is fashion brand Michael Kors, which offers a shopping chatbot on Facebook Messenger. The chatbot determines the user’s location and displays selections available in their region. According to the post, “The shopping assistant makes it easy to browse clothes, interact with the brand, and makes the shopping experience more engaging.” This implementation further demonstrates the power of meeting customers where they are and giving them what they need.
A final example presented in the post is Bol.com, which is a European online store. Its chatbot helps customers with product questions and returns, guides them through the shopping experience in an engaging way, and speaks multiple languages. The post states that this chatbot “can ensure a smooth experience throughout the customer journey.”
A Brief History of a Failed Tech Revolution
When e-commerce chatbots were first introduced, rather than speeding up processes, they actually slowed them down. Chatbots had to search for information and didn’t always find it. When they did, the process was slow, and customers became even more frustrated than they would have been waiting for a human representative to take their call.
Even when chatbots returned responses in a timely manner, they often didn’t answer the questions customers were asking. For example, a customer might have asked “Where is my package right now?” and got back a response saying something like “Your package is on the way.” The customer may have already known the package was on the way, but wanted to know exactly where it was in the process.
Early on, chatbots were also incapable of processing all but the simplest of requests. For example, they could respond to “How can I initiate a return?” but not to “I want to check the status of a return in progress.”
These chatbots further frustrated customers by not being able to retain information already given to it. For instance, if a customer was sent on to a human representative, that person might not be aware of the information the chatbot had already collected.
Even today, chatbots are not seen as entirely trustworthy. According to a recent article on Digital Commerce 360, “Since July 2019, retailers using chatbots in California must identify them to consumers under the Bot Disclosure Act…. Failure to disclose use of a bot can result in a fine of $2,500 per violation.” One of the reasons for this law is that customers want to know who, or what, they are dealing with, so they know what kind of response to expect.
Why E-Commerce Chatbots Fail
While many e-commerce chatbots now in use work relatively well, companies are still searching for resolutions to common issues, such as those listed here:
Language. Chatbots may be programmed to only be able to respond to and in the English language. E-commerce businesses that want to be international must include other languages visitors are likely to use, including Chinese, Spanish, and Arabic. While it’s true that this process complexifies chatbot development, the effort is worth it to ensure all visitors to the site feel they are getting the support they need.
Complexity mismatches. In a recent blog post, Gartner provided a chart that describes the range of complexity that chatbots may be designed to deliver. Low complexity handles simple questions, focused (or transactional) complexity supports conversational interactions, and contextual complexity uses complex models to understand context and personalize interactions. The blog post stressed the importance of pairing the level of complexity with the scenarios it will be used for. A mismatch in either direction can cause frustration and disappointment.
Inadequate design. Some of the issues with chatbots have to do with the fact that we don’t yet have the capabilities to get them to do what we want them to. Specifically, we want them to engage in “natural conversation” as a human would, but the technology simply has not progressed to be able to program chatbots in that way. This is where natural language processing (NLP) will shine, as its main concern is to give computers the ability to communicate through words similarly to humans.
Incremental improvements. Another aspect of the design process that causes e-commerce chatbots to fail is that developers may recognize flaws and fix them, but only incrementally. Thus, if there are high-level issues that should be addressed, the improvements won’t necessarily mend them. Budgetary and other operational considerations may come into play here, as company executives may be reluctant to implement a reworking when minor tweaks are more convenient and less costly.
Too many cooks. Chatbots are designed and programmed by developers, but there are other people involved in these projects, who may have different ideas about what a chatbot should be. In such cases, the end users’ ideas may get lost. As a result, these same users become disenchanted by a service that offers a lot of promises but doesn’t deliver on them.
Hype. Some chatbots’ only failure is the inability to live up to human expectations of what they could become. Even the name we give to this technology, artificial intelligence (that is, the ability to think like a human), is a goal that may still be very far off. Yet, when chatbots don’t behave humanlike enough, many are disappointed.
What Happens When E-Commerce Chatbots Fail
The more e-commerce companies include chatbots in their operations, the higher the risk of not only chatbot failure but also business failure. For example, another Forbes Technology Council article points out, “Chatbots often serve as the first point of contact for a business—and, therefore, play a key role in shaping brand perception.” Yet, “the reality is that customers frequently face obstacles while interacting with chatbots,” negatively impacting their opinion of the brand.
And “negative opinion” can mean different things to different people. For example, if chatbot use is limited to keyboard input, people with limited abilities may be left out of using it entirely. Some chatbots may not understand various languages or dialects. Further, chatbots may not give sufficient contact information for those who aren’t comfortable communicating with an AI interface. And customers may expect the information collected by chatbots to be available if they end up speaking with a human representative, an expectation that is not always met.
Brand loyalty may seem like an amorphous outcome that doesn’t have a big impact on business, but it can. In a highly competitive marketplace, such a seemingly small thing can make the difference between customers’ loyalty or defection. If enough customers have the same experience with this or other small issues and stop doing business with the company, tell others about their experience, or leave negative reviews, companies can lose revenue.
Additionally, if customers lose faith in their ability to communicate with a company through this type of tool, they may demand more human-based solutions that end up costing companies more.
The Future of Chatbots
As the abilities of chatbots, enhanced by AI development services, continue to expand, businesses will be able to use them more expansively, such as to answer more complex questions for customers. Responses from chatbots will become more human-like, and they will be able to recognize subtleties in human communication, such as tone and mood. Images and graphics may be used to support their responses. Companies that save money by hiring fewer humans will be able to save even more as chatbots become more able to take on a wider variety of tasks.
To make the most of chatbots, e-commerce operations should start simply with one function and then add on. They can measure baseline metrics, such as customer satisfaction, revenue, spend per visit, and others to see how much improvement is made over time, then optimize the chatbots for better performance. Beyond customer service, chatbots will be able to help companies create a more personalized experience for customers, giving them an edge over their competition and support for greater business success.
If you enjoyed this, be sure to check out our other AI articles.
