A growing number of companies are using chatbots to market themselves, acquire new customers and drive engagement and sales.
But building and operating a successful chatbot is not a walk in the park. Despite all of the attention lavished on chatbots in the past year, companies face numerous challenges in employing them to create value.
Here are seven of the biggest barriers to chatbot success.
Some use cases are less-than-compelling
While companies like Facebook, Microsoft and Twitter are encouraging companies to build chatbots for platforms they operate, the reality is that some companies are stretching to find compelling use cases for chatbots.
Movie studios, for instance, have created chatbots that allow users to interact with virtual instances of characters from their movies, but the value of these branding and marketing-focused bots is debatable and there’s a strong argument to be made that they are more novelty than anything else.
There is reason to be skeptical about many of the use cases for transactional chatbots, too. While Uber’s Facebook Messenger chatbot, for instance, has been designed to allow users to sign up for Uber and hail rides without leaving Messenger, it’s not clear that Uber customers have a real reason to use the Messenger chatbot instead of the Uber app in the vast majority of cases.
The technology isn’t always good enough
For chatbots that aim to provide users with the ability to interact in pseudo-human way, substantial technology is often required. Conversational chatbots typically often require a combination of artificial intelligence (AI) and natural language processing (NLP).
To assist companies in building conversational chatbots, which is potentially time-consuming and cost-prohibitive, several vendors have created platforms that offer artificial intelligence and natural language processing as a service.
Unfortunately, while AI and NLP technologies are getting better all the time, they’re not perfect and that means that companies take a risk when employing them. Microsoft, one of the biggest technology companies in the world, learned this the hard way when it launched an experimental AI chatbot called Tay.
It didn’t take long for users to learn that Tay, which was designed to learn from its interactions, was incapable of identifying offensive content. Some users took advantage of this to train Tay to make horribly offensive comments – much to the amusement of the internet. As a result, Microsoft was ultimately forced to take the bot offline.
The user experience is often poor
When Facebook launched chatbots for Messenger, many, including industry observers and journalists, logged on to see what all of the fuss was about. By and large, they weren’t impressed. Alex Hern of The Guardian declared, “The bots really suck,” and explained:
“The problems with existing chatbots begin with how they actually work. Almost uniformly, the initial examples of Messenger bots are disastrous: unable to parse any instruction that doesn’t fit their (entirely undocumented) expectations, slow to respond when they are given the correct command, and ultimately useful only for tasks which are trivial to perform through the old apps or websites.”
According to Forrester Research analyst Julie Ask, “It will take five years-plus before these bots are intelligent and can make smart suggestions for us.”
While Facebook’s VP David Marcus earlier this year claimed that much has improved since Facebook launched chatbots on Messenger and stated that the world’s largest social network is “fully invested in [chatbots], and in it for the long haul,” it could prove difficult to convince users to give chatbots more chances down the road if their initial experiences were so poor.
Other technologies could make chatbots obsolete
Improvements in technology and user experience could address chatbots’ biggest shortcomings, but thanks to voice-driven intelligent assistants like Amazon’s Alexa and Apple’s Siri, the market for chatbots could be gone by the time large numbers of chatbots are capable of meeting user expectations and providing sufficient value.
As one user told USA Today, chatbots aren’t especially attractive for ecommerce when compared to voice-driven platforms. “I could use my voice and [make a purchase] in 10 seconds with Amazon Alexa. If I want to order flowers, chatting with a robot takes longer, and isn’t a good experience.”
Efficacy is still uncertain
How effective are chatbots? For chatbots designed to facilitate transactions, such as ecommerce purchases, this is an important question.
At the Facebook Developer Conference back in April, Mark Zuckerberg picked out 1-800-FLOWERS as an example of a retail business successfully using chatbots to drive sales. But while the novelty of “never having to use a phone to order from 1-800-FLOWERS again” may have been attractive initially, the devil is in the details. For example, how do chatbot-driven purchases compare to website-driven purchases in terms of key metrics like customer acquisition cost (CAC), average order value (AOV) and customer lifetime value (CLTV)?
Human assistance is still often preferred
According to a survey conducted by LivePerson, globally, 56% of consumers still prefer to obtain assistance from a live human being rather than a chatbot.
Another survey suggested that 44% of consumers would be open to using a chatbot in customer service interactions “if a company could get the experience right”. However, this is a big if; the suvey also revealed that 88% “expect their chatbot interactions to follow them through their transfer to a live person.”
Delivering on that, for many companies, will obviously be very difficult because it means they will need to have adequate staff available around the clock to manage chatbot hand-offs.
The chatbot market is crowded
Within six months of launching support for chatbots on Facebook Messenger, developers had created more than 30,000 chatbots. While many of these are no more than low-investment experiments, that number highlights the potential challenge companies face in promoting their chatbots.
In such a crowded market, companies are not realistically going to see wide organic adoption of their chatbots without investment of hard and soft marketing dollars. Until the case for that investment can be made, it’s going to be difficult for many companies to market their chatbot sufficiently enough to drive meaningful use.
Agree? Disagree? Let me know what you think in the comments below.