Imagine being able to ask a chatbot questions around-the-clock in as many as 13 different languages. Sounds pretty impressive, right? Well, the KLM Royal Dutch Airlines, in collaboration with Facebook’s Messenger platform, offers precisely this service – and it is a classic example of the proliferation of advanced, artificial intelligence (AI)-powered chatbots in the field of business (in addition to personal mobile AI assistants like Siri, Google Assistant and Cortana). The recent improvements in the standards of machine learning and natural language processing (NLP) have made AI chatbots an integral feature in messaging applications. There are many developers who believe that chatbots can, in fact, replace apps over time. In what follows, we will learn more about the dawn of the chatbots, powered by growth of AI standards:
First things first, let’s understand what a chatbot is and what it is supposed to do. To put it in simple terms, a chatbot is nothing more than a piece of program – designed to initiate and maintain ‘conversation’ with users – simulating the activities of a human customer service representative. A chatbot typically carries on communications on the basis of machine learning, ensuring that the conversations flow in a completely natural manner. Thanks to the underlying AI capabilities, chatbots can ‘understand’ the intent of a person when (s)he sends in a message…and they can respond accordingly.
Rapid growth of global AI markets
By the end of 2017, the value of the worldwide AI market will be around $1.247 million – nearly double of the corresponding figure last year. The growth will gain further momentum as we move towards the next decade – with the AI industry expected to reach $9.55 million in 2021, and a humongous $36.82 million by 2025. The rapidly evolving AI techniques and practices have been instrumental in improving the functionalities of chatbots manifold. Things that were only in the far realms of imagination even half a decade back are actually in place now.
Demand for powerful chatbots
While capabilities and resources for AI-powered chatbots have obviously expanded (the supply-side), the demand for such tools/software has also been increasing from end-users. According to a recent survey, around 7 out of every 10 people wish their initial interactions with businesses to happen via text-based conversations – and chatbots are the ideal tools for that (live chat and messaging both being popular ways). In the bigger scenario too, there is a marked tendency among the present-generation to opt for self-servicing tools. In a chatbot, people can get their queries resolved easily – and without having to depend on anyone else (a business representative, for instance).
Note: The ‘app fatigue’ levels among smartphone-users are also increasing. A vast cross-section of people worldwide prefer to have direct communications with a chatbot, than going through the trouble of downloading, installing and learning about the features/controls of a new mobile application.
4. Moving beyond text-based conversations
Chatbots are far from being limited to only text-based two-way communications (between users and service providers). Things can be made a lot more engaging and interactive with the help of images, emojis, buttons, stickers and even cards. In bots that represent stores or events or a particular product, things like Google Maps driving directions, maps, bios and LinkedIn profiles of concerned individuals can be included. As implementation of artificial intelligence gets more rounded, chatbots become more efficient at bringing all the various types of information that a user might be seeking. The conversations are always contextual…always to the point.
5. A fresh challenge for UX designers
Experts from the field of software and mobile app development feel that 2017 is the year when ‘conversational UX’ really picks up in a big way. The onus is squarely on graphic designers and UX experts to ensure that chatbots are implemented in new applications in a seamless, non-intrusive manner. There are many different types of interfaces and designs that can be created with chatbots – while the conversation types can vary as well. It is a new and exciting challenge…one that app designers need to muster quickly!
6. Exponential popularity of messaging apps
One of the main drivers of the remarkable growth of AI-powered chatbots is the rapidly escalating popularity of messenger apps. In 2016, messaging apps grew 4 times faster than the overall growth rate of mobile apps (44% vs 11%). Both WhatsApp and Facebook Messenger have around 1.2 billion users across the globe – with WeChat and LINE (890 million and 220 million users respectively) taking up the third and fourth spots respectively. Businesses have discovered the benefits of having an active presence on Facebook and the other social platforms, including Pinterest and Instagram. In March, business applications made up 9.88% of the total number of apps in the Apple Store – making it the second-most popular category over there.
7. Efficiency as customer management tools
Chatbots (automated or otherwise) have a big role to play in maintaining the complete ‘customer lifecycle’ and lowering the number of ‘customer-dropoffs’. In this context, the importance of regular notifications and messages sent through these bots has to be highlighted. The AI algorithms can help the software ‘analyse’ and learn from ‘big data’ , facilitate predictive analysis of the collected information, and communicate with users in a personalized, easy manner. In essence, chatbots have the potential to serve as high-efficiency ‘conversational agents’ for businesses.
8. Human assistance should back up AI chatbots
Irrespective of how nuanced AI technologies have become, and what range of functions that chatbots can serve – the latter should never be a ‘replacement’ or a ‘substitute’ of the human touch. There can always be a query that a chatbot might fail to understand, or can be answered better by a human customer service executive. While creating the underlying algos, developers need to be on the lookout to provide opportunities where a human can step in an ongoing conversation between the chatbot and a prospective customer.
Note: In general, questions asked by chatbots should be choice-based (e.g., which size of shirt a customer wants to buy, along with the available options). Questions that invite open-ended answers are not ideal for chatbot conversations.
Multiple tasks handled by the same platform
Say, you wish to buy something online. You have to browse through the concerned category at the online shopping store, arrive at a final purchase decision, place the order, and make the payment through a secure gateway. After that, you would also probably like to receive notifications related to the shipping status from time to time. Traditionally, all these tasks had to be conducted separately – something that has been resolved by the cutting-edge chatbots. On them, the different tasks associated with making a purchase can all be carried out on the same platform, within a single conversation. Unlike simple mobile shopping apps, chatbots do away with the need for following different message threads while doing a purchase.
The cost advantage
Developing a chatbot for business makes sense from two perspectives. Firstly, by enabling direct communication with people and ensuring complete personalization, chatbots bring down the total expenses associated with customer services. Secondly, and equally importantly, the average costs for making a chatbot application are also steadily going down – thanks to the availability of an excellent array of free development frameworks (released by Google, Microsoft, IBM, Facebook). What’s more – chatbots are server-side apps that require very basic UI elements. Hence, they can be created more quickly than an average smartphone application.
Support from the big players
2016 was a landmark year in the growth of messenger chatbots. Facebook Messenger announced the support for bots in April – barely a month after the much-publicized Bot framework had been announced by Microsoft. The total count of ‘approved bots’ – for platforms like Telegram, Slack, FB Messenger and Skype – has gone up significantly, pushing up the total user base of messenger bots to well above 1 billion individuals. With artificial intelligence continuing to push boundaries, many new types of chatbots – utilizing the latest opportunities – are likely to be launched over the next few quarters.
Note: Voice recognition is also fast emerging as an important feature of AI-powered chatbots. You can check out the impact of speech recognition technology on mobile apps here.
12. The ethical considerations
Fake news, modified algorithms, spammy notifications – all of these can compromise the quality and usability of a chatbot application. A customer can typically share sensitive, personal information during a conversation – and the business has to make sure that such information can never be accessed by unauthorized third-parties. In addition, community managers (social media experts) and customer service professionals have to closely collaborate while working with a chatbot. In general, customers expect the two roles to be performed by one and the same person…and it is important to ensure that this perception is maintained.
Chatbots have come a long way since the arrival of ELIZA – the MIT program – way back in 1966. In their present form, they can offer personalized, real-time information to individuals on the one hand…and collect & analyse data (data mining) on the other. Artificial intelligence has been instrumental in making chatbot conversations more effective and streamlined – with automated suggestions also being a big help. Not surprisingly, many leading business – from Amtrak and Disney, to Mattel and Taco Bell - have joined in on the chatbot bandwagon. Front-end response systems of businesses (maintained mostly by website and mobile applications till now) are all set to undergo an overhaul – and AI-based chatbots are going to be right at the core of this revolution.
Latest posts by Hussain Fakhruddin (see all)
- AI-as-a-Service: All That You Need To Know - October 30, 2018
- Top 12 Trends In Digital Transformation To Watch Out For In 2019 - October 23, 2018
- Machine Learning in 2019: Tracing The Artificial Intelligence Growth Path - October 9, 2018