Upcoming WebinarsMore Webinars
Alan Dix explores AI and Social Justice
Author of Introduction to Artificial Intelligence
AI Business is part of the Informa Tech Division of Informa PLC
This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.
by Jiaqi Pan
BARCELONA - There’s no denying that artificial intelligence (AI) is changing what is possible when it comes to tech applications. Machines that learn is the stuff of science fiction - so it should come as no surprise that many startups today are attempting to leverage the power of AI into their operations. However, the majority of these startups are not asking themselves whether or not their operations need such tech in the first place.
The world of chatbots is the perfect example. Companies are supercharging customer engagement tools which simply do not need to be supercharged. The resulting user tools are undoubtedly high-tech - but also expensive and generally inefficient. Biggest does not always mean best, and smartest does not always mean most logical.
In fact, a growing number of businesses are bucking the AI trend to keep their chatbots “dumb” - with a simple conversational approach that flows naturally and converts accordingly. Let’s explore why - when it comes to chatbots - it pays to keep it simple, stupid.
Chatbots are conversational interfaces which in recent years have become almost ubiquitous across the World Wide Web. Along with their voice-enabled cousins like Alexa and Siri?, chatbots are conversational interfaces designed to answer any and all customer inquiries. This requires conversational content programmed in conversational units, also known as question-and-answer pairs. To do this well is not as easy as it may appear.
Chatbots - when done right - can seamlessly answer questions and respond to commands from consumers. In the majority of cases, chatbots interact in a conversational manner like a digital content tour guide and direct customers to what they are looking for. However, customers know when a chatbot has been poorly implemented as they frustrate, confuse, and ultimately bore.
Nonetheless, the adoption of these online customer helpers continues to grow year-on-year. In fact, 80 percent of businesses are planning to deploy some type of chatbot by 2020 - and this is for a wide range of uses. The Content Wrangler and Precision Content Authoring Solutions survey found that 65 percent of respondents will use chatbots to deliver customer support and technical documentation content to those who need it. Similarly, 30 percent said they plan to use chatbots to deliver guided customer journeys designed specifically to help navigate a set of content to achieve a goal.
With chatbots gaining popularity, it only makes sense that developers are attempting to make them better too. Enter artificial intelligence (AI) - the tech trend which is undoubtedly making chatbots smarter while simultaneously detracting from the user experience.
AI has progressed exponentially in recent years to become one of the most exciting tech possibilities in contemporary development. With great power comes great hype - and outfits in countless industries are attempting to leverage the technology even when it is unnecessary. For example, a survey of more than 2800 self-described “AI startups” found that about 40 percent could not give any evidence to support their machine learning claims.
This is clearly occurring in the world of chatbots. Developers are unleashing AI onto tools which are often better left rudimentary, transforming the chatbot into the “personal assistant”. For example, Google’s Duplex voice AI is heavily trained to perform very specific and usually formulaic tasks — such as booking a hair appointment or a restaurant. Basic function demands basic form, and the AI craze sweeping the world is actually damaging the usefulness of chatbots.
In the end, businesses pay more for a chatbot with AI when it simply does not require AI. The majority of these bots often realize poor adoption rates, high user frustration, and limited benefit to the corporation. In fact, they typically hurt an enterprise. Costs go up while customer satisfaction goes down. It all begs the question: do we need to dumb down our smart chatbots?
Common sense will say that smarter is always better - but this is not the case when it comes to the chatbot. Remember, these bots are created to generate leads and assist with customer inquiries. More often than not, this can be achieved with an efficient yet AI-less conversational tool.
The fact of the matter is that customers do not care how powerful the chatbot is. They care about how helpful and useful it is - and that is not synonymous with AI. Rather, a well-designed dumb chatbot can be much more effective for the user than an ill-advised smart chatbot.
It all comes down to design and purpose. AI chatbots have a huge gap between user expectation and reality. No matter how well-trained our chatbots are, users want the experience of speaking with another person.
Conversational chatbots are quickly gaining traction in offering what both the client and customer desire: a basic yet intuitive design. In the Content Wrangler survey, 73 percent of companies that have implemented chatbots say their content is written in a conversational style. A further 10 percent said their content was not written conversationally, but that it should be.
This is because it works. Chatbots are designed to chat and answer queries, something that simply does not require high-end technological capabilities. Further, the better chatbots of today already offer statistical and demographic preference prediction and database parameterization such as name, location, appointments and prior purchases.
A turning point is on the horizon as businesses realize their high-tech AI tools are often less efficient than their dumb ones. AI may be game-changing, but it should not chatbot-changing. The opposite is true - chatbots need to keep it simple to continue finding success with lead generation and user assistance.
If it ain’t broke, don’t fix it - and especially don’t fix it with unnecessary AI.
Jiaqi Pan is CEO of Landbot.io
Author of Introduction to Artificial Intelligence