7 minsApril 10, 2019
Whenever there is a forward-looking article dealing with technology, it usually makes relatively vague predictions based on what we understand about an upcoming technology, at the given moment. Usually this results in a rather simplistic exercise
in extrapolation rather than true predictions. However, this one is about a piece of technology that has already become quite popular and something a lot of us depend on regularly, whether we know it or not – we’re talking about
Much has been said about things like voice search, augmented reality and even smart voice assistants, but all these are still trying to overcome technological limitations that prevent them from becoming truly useful.
Chatbots, on the other hand, are technically simpler to implement, at least simple ones like the kind that dish out canned responses on Facebook pages. They rely on the textual aspect of the internet, mostly on instant messaging platforms that
we’re already too familiar with. Despite the simplicity, they lie at an interesting intersection of automation, artificial intelligence and the evolving nature of the World Wide Web. Smart chatbots which understand context and are self-learning
are within the realm of possibility but still far away and limited by technology. So, before we try to analyse them further, let’s take a look at what chatbots exactly are.
What are chatbots?
A bot is a piece of software that automates a task for you. A chatbot is a bot that does this through text messaging. The task being automated could simply be a conversation, or even something like ordering a pizza. The program powering the chatbot
itself is predetermined and used to be something that only developers with the correct knowhow could build.
However, today, with the existence of multiple drag-n-drop platforms that allow newbies to build the bots, chatbots have been truly democratised and put to use by anyone who sees the need. It is difficult not to see it, since chatbots today are
much more than just conversational engines. They can handle speech-to-text, computer vision, translations, text analysis and more!
As of May last year, there were more than 300,000 chatbots on Facebook Messenger alone, and more than 150,000 accounts registered on wit.ai, a natural language processing platform intended for developers. However, let's not just rely on global
Homegrown Indian startup Niki.ai, as of June 2018, was providing a host of transactional services to over 2 million customers. A number of Indian companies, tech or otherwise, are deploying chatbots to serve various purposes like customer service,
answering frequently asked questions, order placement and more. All of this is coming on the back of the growing Internet citizenship of the country. According to IAMAI, India will have an online population of 500 million by 2022, while some
reports say that we're adding 10 million new netizens every month. This unprecedented growth is backed by cheap smartphones and cheaper data. It forms the ideal base for chatbots to takeoff. This is exactly what they’re doing now.
Take the banking sector. Axis Aha, the conversational digital assistant from Axis bank, can not only answer customer queries via voice and text, it
can also help them execute actual transactions and actions like blocking a card, recharges and more. It can intelligently route to an agent in case it is unable to answer the query and there is a continuous learning and feedback mechanism
which ensures that the bot responses are accurate and updated.
Axis Aha! is just about starting on its journey and has already answered over 8.2 million FAQs since its launch, with over 87% accuracy. And things are only about to get better.
The future possibilities
According to a study by Juniper Research, chatbots will redefine the customer service industry, with banking leading the pack. The report predicts the success rate of chatbot interactions to cross 90% by 2022 in this sector. Which means that more
than 90% of conversations that banking chatbots will have in 2022 will reach successful resolutions without human intervention. Therefore, something like Axis Aha! will gain a lot of importance
and capabilities over the next few years. Some such possibilities are as follows:
- Contextual awareness: Some of the most widely used conversational agents are found today as virtual assistants. We're talking about Siri, Alexa, Google Assistant and the likes. As 5G becomes more prevalent, chatbots will begin
to grab data from all kinds of sensors and smart devices around a user and offload most of the processing for this large amount of data onto servers. This will be to bring you a coherent, useful virtual assistant experience – smart
assistants finally getting smarter!
- Sentiment awareness: Apart from contextual data and its better identification, bots will also gain the ability to tell your mood depending on how you’re texting it. This already does exist in rudimentary form in terms
of financial institutions using NLP (Natural Language Processing) to gauge market sentiment by parsing public posts. But this is expected to get much better and nuanced at identifying sentiment at an individual level with the actions you’re
trying to undertake.
- Bots talking to bots: As more services deploy chatbots to handle operations, we could be looking at a situation where one chatbot needs to talk to another to get something done. For instance, your food app chatbot talking
to your banking app chatbot to process the payment for your order, without you having said anything more than ‘I want pizza for dinner’.
- Lifelike conversations: Quite recently, Elon Musk-backed OpenAI announced the development of an AI writing program that could analyse text on the internet as a source and write articles that were difficult to find flaws with.
Developments like this hold a lot of promise when it comes to chatbots getting more lifelike in their conversational skills. Giving such AI the right dataset to learn from could allow the creation of software-backed digital customer care
agents that are indistinguishable from the real thing!
- Voice takes over: A major issue with using a voice-bot has been how unnatural it feels. To actually get something done, it feels like one is talking to an IVR system. However, if Google’s Duplex demonstration from last
year and follow-up user accounts are to be believed, we’re nearly on the verge of voice bots talking like real people, without anyone knowing any better. This could be a huge leap for chatbots as voice is always a more comfortable
medium than text. Chatbots will of course eventually evolve to employ voice.
- Single-unified interface: One of the main reasons that chatbots are being touted as the ‘replacement’ for apps is because of just how many apps are out there. However, just like how apps didn’t replace websites,
chatbots and apps will coexist. Although, you’ll be using a lot fewer apps as more services start offering chatbots. Perhaps in the future a single unified interface, something like Facebook Messenger or WhatsApp, will be used to
deploy most chatbot services.
The web is already redefining how we interact with chatbots. Search engines like Bing have started to show bots on the search engine result pages (SERPs) directly. In Apple map results, you can directly text a local business to get a response
from a chatbot. Before you realise, chatbots will surround you as ubiquitously as apps do today. Will you be ready to have a word? Try talking to the Axis Aha! chatbot right now by clicking here and experience it for yourself.
Disclaimer: This article has been authored by Siddharth Parwatay, a Mumbai based independent tech-journalist, editor, and content-creator. Axis Bank doesn't influence any views of the author in any way. Axis Bank and/or the author shall not be responsible for any direct / indirect loss or liability incurred by the reader for taking any financial decisions based on the contents and information. Please consult your financial advisor before making any financial decision.