Building The Branded Chatbots Of The Future
A startup is creating intelligent and responsive human-like AI to serve an ever-expanding range of roles
Chatbots have been gaining exposure ever since Zuckerberg announced his grand plan to integrate them into Facebook’s Messenger platform. Employing these ‘chat robots’ will make for a more dynamic system, one in which users will be able to speak directly with businesses the same way they do with their friends. The implementation is an attempt to secure Messenger’s standing as a central hub for all digital interactions, and by onboarding brands onto its platform, its slated to become one of the largest B2C channels out there. At the same time, participating companies get access to a user base of some 900 million users as of April. But chatbots don’t merely exist within the realm of Facebook, and one company is betting on their implementation outside of Messenger. As a bot-maker, Kore creates intelligent, human-like AI for major partners like Salesforce, The New York Times, eBay, Dropbox, CNN, LinkedIn, ESPN, Business Insider, TED, and many more.
Evidently well-versed in bridging the company/user divide, we sat down with Kore’s Global Head of Solution Management Amit Aghara to discuss what it takes to build a bot, how the practice is evolving and how their implementation will shape virtually every industry in the near future.
“Messaging is an interesting paradigm in how people communicate,” Aghara stated. “It’s natural and intuitive, [the user experience is the same across the board] and you don’t have to force it onto people—the tendency to use it is already ingrained. To make an effective bot, you’re going to have to adhere to the rules set in place by the messenger; the way the bot interacts is limited by the technical nuances of the platform it’s in, and as a primary communication methodology everyone will want it to be quick, seamless and conversational.”
For this reason, Aghara explains that businesses are better off starting with a chatbot outside of Facebook Messenger. With the pipes already laid out, porting the bot from Messenger to other platforms becomes a complicated chore, as the bot becomes inflexible once implemented.
Instead, he advises businesses new to chatbots to build a more basic bot capable of integrating smoothly into all interfaces, and only adding features after the bots have already been implemented into their respective locations.
In this way, the inclusion of a company’s chatbots across various channels will be consistent and omnipresent. Rather than a fragmented user experience, Kore’s solution comes to life across all outlets where a brand’s bots live—if a bot learns something about the user on one platform, it learns it in all of them. And on that note, Aghara adds:
“We already know that bots need to be able to communicate with pure, natural language, which is where natural language processing comes in to make the interactions customized and actionable for the user. But what’s less obvious is that it shouldn’t end with your response; the bot needs to learn your behaviors and respond accordingly. It needs to build contextual awareness based on the metadata you provide: this is what will truly set differing experiences apart from one another.”
Kore’s ultimate vision is a perpetually learning, universal chatbot across all channels, including ones like Amazon Alexa.
Aghara explains that support for such systems is coming, and that it shouldn’t be too difficult to implement since Alexa converts speech to text for processing as is. Coupled with the release of the Alexa Skills Kit (ASK), fleshing out these bots won’t be difficult for Kore.
The learnings from Kore are threefold: build platform-agnostic bots to start with, implement learnability into the core functionality so that the bot can become contextually aware, and most importantly, make sure your chatbot speaks ‘human.’ Remember that chatbots exist to be spoken with, and therefore take commands through text and speech. If you’ve ever spoken to a robot, you can imagine how stiff the conversation can feel—removing that rigidity (without sacrificing the utility of the chatbot) will go a long way in ensuring consumers feel a real connection with your brand.