>Benjamin Gaines

Bottom Line Up Front

The Good

  1. Design chatbots to keep customers in control

  2. Set high standards for minimum viable products

  3. Have a data strategy in place to use gleaned data

The Bad

  1. Don’t let your chatbot be creepy

  2. Avoid thinking of a chatbot as a stand-alone “IT Project

The Ugly

  1. Chatbots are incapable of empathy

  2. Chatbots are still in their early stages of development

Intro

Chatbots - computer programs designed to simulate human-like conversations through artificial intelligence (AI) - are one of the of the most popular applications of AI in business right now. Research indicates customers are increasingly comfortable speaking with chatbots. In fact, some customers actually prefer speaking to a bot over speaking to a human. Accordingly, chatbots are a frequent topic of discussion with our clients.

The business case for chatbots can be quite compelling. They can drive growth by improving customer experience while also reducing cost by creating efficiencies. In addition, chatbots that use AI technology such as natural language processing and natural language generation are quite advanced and commonly used. Several vendors offer out-of-the-box solutions for these technologies.

There are, however, many factors that need to be considered when designing and implementing chatbot solutions. If not done properly, chatbots can irritate customers, damage the company’s brand, and potentially add a tremendous burden to development teams. So, what differentiates companies who succeed with chatbots? What are traits of successful bots (and bad ones)? In this post, I share a brief point of view on chatbots. In doing so, I divide my thoughts into three parts: the good, the bad, and the ugly.

The Good

  1. Design chatbots to keep customers in control. Quality chatbots inform and empower customers. They remind customers they are interacting with a bot, not a human, but still assure the customer they are in charge. The most important way they do this is by immediately responding to demands or directing them to a human representative who can. When building and testing chatbots, therefore, do more than simply just checking the code and functionality. Check the experience. Always ask, “Do I feel in control right now?” If the answer is no, it doesn’t matter if the code works; the chatbot is on the wrong track.

  2. Set high standards for minimum viable products. The typical approach to developing tools like chatbots includes building the smallest working product (“Minimum Viable Product”), launching that product, and then improving iteratively with feedback about it. There is, however, an inherent hazard in launching a minimum viable product that stands to annoy customers. Not only could it irritate customers, it can put the development team in a cycle of responding to negative feedback. Instead of asking what customers need, and how can that need be satisfied, development teams will be racing to patch problems with a chatbot that irks customers. That’s a spiral to avoid. To do so, set very high standards for a MVP… even if it delays the launch a little. Your customers will thank you.

  3. Have a data strategy in place to use gleaned data. Chatbots collect a tremendous amount of customer data. Gaining actionable business intelligence from that data is key. Firms should therefore have a solid data strategy in place before chatbots are even designed. Similar to the aforementioned guidance about asking if the chatbot makes the user feel in control, a chatbot should constantly help you learn more about your customer.

The Bad

  1. Don’t let your chatbot be creepy. Keeping in line with the “keep the customer in control” guidance, don’t use chatbots to pursue unsolicited follow-ups with customers. When customers indicate they want more information on a given topic, chatbots are an excellent way to facilitate their requests. But when a customer hangs up, or clicks on something else, do not use chatbots to generate a spam email, text, phone call, or targeted ads for issues unrelated to their inquiry. Seemingly random bot-generated follow-ups can be considered creepy. They startle customers and actually work against any opportunities businesses are trying to create. If a chatbot uncovers a potential opportunity or lead unrelated to the customers initial inquiry, have an empathetic, relatable customer representative do the follow up.

  2. Avoid thinking of a chatbot as a stand-alone “IT Project.” Customers use their interactions with a chatbot to form an opinion on the entire company. The stakes are high. If customers are shocked by how cool and seamless a chatbot is (Siri, for example), they’ll think the whole company has high standards. Conversely, when they come across a poorly formed chatbot, they’ll think the company is just trying to use chatbots to cut costs and fire people. Therefore, don’t consider the creation of a chatbot to be a simple, stand-alone, “IT project.” Chatbots are a multi-team effort that should keep the customer’s needs at the center of the initiative. Every department involved needs to roll up their sleeves and give it their all. If you decide to build a chatbot, plan on adding resources, and likely even losing the time from some of your top people. This is not an initiative to give a half effort, but a journey to impress your customers and provide a more meaningful experience.

The Ugly

  1. Chatbots are incapable of empathy. The vast majority of customer interactions in business services are for routine tasks wherein chatbots are a fantastic option. There are, however, still some interactions which demand a level of empathy chatbots simply cannot deliver. (Death of loved ones, catastrophic loss of property, divorce, bankruptcy, etc.) In these situations, companies can have lasting impressions with customers by meeting their needs and making them feel cared for. Chatbots aren’t perfectly suited for this. Wise companies recognize this paradigm and increase initiatives to drive human-based, emotional understanding of their customers along with the development of bots.

  2. Chatbots are still in their early stages of development. Though the business cases for chatbots can be compelling, return on investment (ROI) is not guaranteed. Chatbots are still a relatively new phenomenon. Each companies chatbot journey varies slightly. So too does their customers responses. There are a variety of company specific situations that could alter the ROI for a given chatbot. Accordingly, it might not be possible to benchmark the ROI of your chatbot against a competitor.