Why Chatbot Fails Happen: Surprising Insights Revealed

Chatbot fails can frustrate users and damage brand trust, but what really causes these missteps? Understanding the hidden pitfalls can help businesses create smarter, more effective bots.

But there’s no need to panic because you can rest assured that chatbots aren’t bloodthirsty androids (yet) — they just want to help answer FAQs, speed up sales processes and lighten the load of customer support representatives.

A report from Forrester Research showed that more than half of the companies surveyed stated that they either had a chatbot system in place, or were planning on developing such a system within the next twelve months.

While chatbots can make some tasks easier for both vendors and prospective customers, these systems are only as good as the user experience they provide. This article will examine how a chatbot works, the primary mistakes companies encounter when they develop their chatbots, and ways to prevent those mistakes from creating an unpleasant experience for the user (and an unprofitable one for the company).

What is a chatbot?

According to Chatbots Magazine, a chatbot is “a service, powered by rules and sometimes (by) artificial intelligence, that you interact with via a chat interface.”

A chatbot that relies on rules can only accept a limited number of inquiry types, and can only respond in limited ways. A chatbot that uses “artificial intelligence” employs sophisticated algorithms, such as Natural Language Processing, to handle user requests.

Regardless of which system powers the chatbot, users will quickly abandon sites that employ chatbots that provide an awkward and difficult user experience. According to Matt Hartman, Director of Seed Investments at Betaworks,

“The difficulty in building a chatbot is less a technical one and more an issue of user experience. The most successful bots will be the ones that users want to come back to regularly and that provide consistent value.”

[Chatbot fail #1] Misaligned messaging

Your chatbot says a lot about your brand, literally.

If you’re in the business of selling children’s toys, your chatbot needs to reflect the tone and personality of your brand, which should be casual and bubbly. But that same causal approach won’t wash so well if you’re in the B2B aviation industry.

Putting tone to one side, chatbots also need to deliver clear, concise, and conversational responses. The best way to keep a user engaged is to keep the chatbot's messages short, simple, and to the point. While this approach may preclude a chatbot from accomplishing more sophisticated tasks, the emphasis on delivering simple and direct answers will make the process easier for both chatbot developers and the users who engage with them. Both sides are often better served by having the chatbot accomplish a few simple tasks flawlessly, rather than attempting to do too much and failing miserably.

An example to follow: Sephora’s on-brand chatbot customer experience

Sephora’s chatbot is a good example to note. They’re a professional makeup brand, and you get that feeling when you talk to their bot. It greets you formally and then cuts to the chase by giving you three options to move you in the right direction (book a makeover, share feedback or talk to an agent). As you will come to see in the other chatbot experience examples we have lined up, this strategy isn’t always the case. In fact, Sephora has deliberately avoided emojis and whacky GIFs in order to stay on brand.

[Chatbot fail #2] Conversational dead ends

An aspect of delivering clear and concise messages with a chatbot includes ensuring that the conversation between the bot and the user doesn't stall into an unproductive sinkhole. When the chatbot fails to offer the user any useful options, the user can get lost. When users encounter a “dead end” with the chatbot, they leave the site and the brand ends up paying for a chatbot platform that does nothing but kills conversion rates.

However, when a chatbot offers the user clear options, the user will stay engaged and move further along the conversion process. For instance, instead of ending a conversation with, “I hope I was helpful!”, the chatbot can ask, “I hope I was helpful! Which product lines did you want to know more about today?” or “Which of our services did you want to ask me about?” This approach ensures that the user gets the answers they need and helps them with their purchasing decisions.

An example to follow: Marriott Hotel’s talkative chatbot customer experience

The Facebook Messenger chatbot from Marriott Hotels embodies a good conversation strategy. When you want to stay in a place or at a time when there are no rooms available, the chatbot doesn’t just fob you off with a “sorry” message, nor does it ask you to check their website. Instead, it keeps you engaged in the conversation by giving you further options.

[Chatbot fail #3] Lack of personality

When it comes to any type of interaction, humans want to feel like they're engaged in a dialogue with someone — or something — that can relate to them. While the idea of adding a “personality” to a chatbot may sound strange, it does help customers engage with the bot and the site. Just because users know that they're chatting with a machine, doesn't mean that they should feel like they are.

Thankfully, the process of giving your chatbot a personality isn’t anything super rocket science-y. You could give the chatbot its own name, some dialogue mannerisms and even some baked-in comments that can be sent to the user depending on the time of day, season and product they’re asking about. All these minute details help make your chatbot more relatable.

An example to follow: LEGO’s personable chatbot customer experience

LEGO is one brand that has done well to inject personality into its bot. Their bot greets you with a robot emoji and GIF, and continues to make LEGO-related quips throughout the conversational experience.

[Chatbot fail #4] Going in strategy-less

Many companies make the mistake of deploying new technologies strictly for the sake of deploying them. From websites to mobile apps to email newsletters, technologies should only be used if they can further the brand’s missions. This truism also holds for chatbots. A clear strategy involves an understanding of the bot's purpose, its key functionalities, and the ways in which it will serve the customer.

Just as sales staffers and customer service representatives have clear strategies in place to lead customers through the conversion process or address a customer's concerns, chatbot developers should have the same chatbot strategies in mind when mapping conversations between the bot and the user. While the objective for nearly any business strategy is to convert users into customers, the chatbot strategies must strike a balance between steering the customer toward a purchase and coming across as too aggressive or “sales-y”. Users will turn off or report bots that send constant sales spam messages, which can cause the bot's listing to be removed and cost the company in lost opportunities.

An example to follow: eBay’s sales-focused chatbot customer experience

eBay’s Shopbot is a good example to follow here. The strategy is simple — to help shoppers find great deals. But the execution of that strategy is excellent. It starts off by offering the user a range of trending deals before asking the user if they’re looking for a specific product.

I typed in “iPhones”, and the bot immediately asked which network I preferred. I was then presented with a range of iPhones, before being asked to specify my budget. The conversation goes on and on until the bot presents you with a highly targeted range of products that fit the bill.

Make your chatbot count

Chatbots are quickly becoming a prevalent part of the user experience., but you don’t want to be the business that deploys chatbots because they’re fun or charming. You want your chatbot to deliver results — like a return on your investment.

As the presence of chatbots increases, so will users' expectations of what chatbots can deliver. While chatbots do not need to be all things to all users, they do need to meet a user's specific needs based on their requests. A major component to the success of a chatbot lies less in its ability to deliver the builders' message, and more on its ability to “listen” and deliver the message the customer needs. So, study the fails above, and make sure your chatbot wins.