subscribe: Daily Newsletter

 

Why the future is bright for chatbots

0 comments

If you have been following technology events over the past couple of months, you may well think that chatbots have taken over the world, writes Stephen Ball, senior vice-president: Europe and Africa at Aspect Software.
In April 2016, Facebook announced chatbot capabilities on their services, specifically targeting customer engagement. Since then, several other developers claimed to have made chatbots for the likes of WhatsApp, Slack and other messaging services. Facebook even announced that “tens of thousands of developers” are building software to take advantage of this function.
However, there have been mixed opinions in various media outlets and in customer engagement circles about the usefulness of chatbots. The Guardian notably published a damning review of the early entrants, such as CNN and Hi Poncho; neither of which supported a natural, enriching experience.
Early examples of chatbots have been underwhelming and quite frustrating. If a chatbot delivers a remarkable number of notifications and does not provide an easy ‘off’ switch, logically consumers may claim that this kind of ‘pushy’ automation merely benefits the brand. For example, when a user cannot unsubscribe from notifications because a chatbot does not recognise natural language (i.e., “I want to unsubscribe” cannot be understood), developers must acknowledge the serious user experience issues with their products.
The Microsoft chatbot test account, Tay, is a good case study detailing the challenge of making the best use of natural language learning. Anyone could chat with Tay via social media, with Tay building her lexicon and understanding of context based on her interactions. The aim was to have Tay appear as natural as possible, emulating the language that customers use in order to avoid sounding robotic or stagnant. Unfortunately, Microsoft didn’t reckon on the abundance of troublemakers and trolls online and it wasn’t long before Tay was bleating obscenities and Microsoft took her offline.
Given all of this, brands may be thinking: Can chatbots be anything more than ‘spammy’ marketing tools? What is the value that chatbots can bring to the overall customer experience?
There’s clearly a better scenario for brands once they decide to add a chatbot to their customer service repertoire. Organisations must identify where a chatbot might best sit within the customer journey and be able to act as a valuable addition to the human advisor team. Nobody wants a robot designed to chuck out FAQ answers, and even then only once it’s been given the ‘correct’ (and in many cases, non-obvious) commands by the user.
Unlike Tay, chatbots can greatly improve customer interactions, and need not be frustrating, alienating or rude, even if they are ‘fed’ curse words and insults.
The main benefits of chatbots stem from the growing demand from consumers for new ways to solve their problems and interact with brands and services. Back in 2012, Ofcom noted that texting and instant messaging have become the most common forms of communication, far ahead of talking over the phone. This applies to interactions with businesses as well as other consumers; with customers looking for solutions to their problems away from the traditional voice chat with a contact centre representative, including self-service and text-over-talk options.
This is part of a growing ‘millennial-mindset’, with consumers relying on their mobile devices for the majority of their communications.
Chatbots, when used correctly, can be perfectly positioned to meet this demand in customer engagement environments. Even if a chatbot can only understand simple requests (“send receipt”, “broken phone” etc.), it is still enabling customers to begin to solve their issues without needing to speak to anyone. This is true even if the bot isn’t sophisticated enough to resolve their issue entirely, since sophisticated chatbots can hand over to a human advisor in a seamless way.
However, were a bot able to understand natural language, then more nuanced customer service issues are more likely to be solved entirely through self-service. All communication can take place over SMS, instant messenger or email, with the customer talking effortlessly with the chatbot and the software responding to their questions appropriately.
Of course, agents may still need to be available for the most complex of queries, but the majority of interactions through chatbots could be solved quickly and easily.
Some companies are already employing chatbots with this level of sophistication, such as Edwardian Hotels London (EHL). Earlier this year the hotel chain introduced its guests to ‘Edward’, an artificially intelligent, SMS-based virtual concierge. The chatbot enables guests to check and request amenities (such as extra towels or room service), get information about local bars and restaurants and even express complaints, all over text. Edward also has natural language capabilities, being fully able to understand complex demands in a conversational tone, such as “where is the best place to eat Chinese food” or “how much is a taxi into town?”
Clearly, despite the shortcomings of some of the initial offerings to market, chatbots have the potential to contribute and enhance the customer experience for brands. Far from deploying the technology as an ‘extra’ or novelty item, chatbots should act as a development in self-service and add value to the customer experience. Adopting a customer contact channel must have a business case; from customer demand to saving time and money for either the customer or the business (or both), chatbots appear to be meeting that need. In order to move on and reap the benefits of the technology, we must make chatbots that truly feel like having a ‘chat’ with a real human, and one that is knowledgeable, quick and in many applications, secure.
Brands want to be where their consumers are, and without a doubt that is on their mobile, texting, messaging and posting. The proper use of chatbots could mean that a consumer interacting with their favourite brands in this way eventually becomes as natural as texting a friend.