Chatbots or conversational AI has come a long way from the early days. This was the time when the technology struggled to understand a person due to their accent or because the chatbot had only a handful of responses to offer, none of which met a specific query. However, with development in the field of Natural Language Processing (NLP), and AI, chatbots are not only becoming multi-linguists, but also adept at handling complex or multistep tasks such as checking daily schedule, making suggestions on the books you could read or making a customer understand why his latest bill is higher than the normal.
The global market for conversational AI is growing at a CAGR of 22 percent and is expected to be worth over $14 billion. What this indicates is that chatbots (the most popular conversational AI tools currently) are only going to grow faster in the years ahead.
As we saw during this year, this rapid growth is being enabled by innovations and the expansion of language coverage offered by these tools. Despite the global usage of English, it is still not the first language for majority of the world’s population.This is more so in a country like India where nearly 90 percent of people prefer to speak in native languages.
Thanks to Cloud, the deployment of modern conversational AI across diverse devices and platforms has become much easier and faster. There is a significant reduction in set-up challenges, the hours of training required and the cost of maintenance as well.
According to Gartner, by the end of 2022, about 70 percent of executives will regularly use conversational AI tools including chatbots, virtual agents or voice assistants etc. Some other studies and surveys that reveal 9 out of 10 brands reported faster complaint resolution when they used conversational AI.
In 2021, we have seen a lot of conversational AI and AI voice analytics deployment in various operational areas such as customer service. It is not that the technology has achieved its true potential, but rather that there are some innovative conversational AI brands using advancements in NLP and analytics to make conversational bots more useful in diverse arenas.
One of the biggest trends of voice AI adoption has been in the area of training process improvement. For any industry, it is important to train AI on industry specific content. In a typical call center setting, the voice AI or chatbots will be trained by using data such as live calls or chat transcripts, call recordings and business FAQ documents. This rule-based or manual training is time consuming, and deployment takes several months.
If you try to expedite the activation, then there could be inadequate understanding of the scenarios and less than satisfactory responses for the users. This would defeat the whole purpose of investing in such platforms. This is where innovations to automate and accelerate the training process have been a highlight of 2021.
Usage of not just inputs, but also sentiment and intent analysis has led to conversational AI delivering more relevant and qualitative responses. In India, the ability of a chatbot to understand and respond in a native language such as Hindi, Bengali or Tamil is of extreme value. It helps in establishing a stronger rapport between the users and the machines. Further, we have seen how NLP has made machines sound a lot more human. It is the deployment of such tools that is a heartening trend. It gained momentum in 2021 in not only customer service, but also in areas like healthcare.
Personalized AI for superior performance
As online shoppers would vouch, AI is becoming smarter and personalized. From analysing a customer’s online behaviour to their purchase history, conversational AI tools are able to make better recommendations. It could be a suggestion related to switching to a ‘data heavy’ plan for a telecom operator, to suggesting what one could have for dinner (based on the frequency and type of food ordered online). The bottom line is that conversational AI is enhancing user experience, reducing cost of conversion, and leading to less escalation of routine queries at contact centers courtesy of the increased ability to deliver more empathetic and engaging responses.
Predictions for 2022
Going forward, we will see AI being deployed in more complex scenarios. While chatbots are the dominant form of conversational AI, voice-enabled solutions will now gain ground. For instance, the conventional text-based search is already being phased out by voice search, and a similar Zero-UI concept of computers that talk back is going to be the new norm.
With greater NLP and real-time analytics input, training spans will become shorter, and computing requirement will be significantly reduced. Some of the best Indian voice AI systems are capable of working in very low network coverage areas and consume up to ten times less compute power than a typical system. This is going to enable usage of such technology in almost every domain with greater accuracy.
We might not yet be at the magical level where R2D2 or C3PO can flawlessly converse with everyone, but we are definitely getting closer each day!
(Tapan Barman is Co-founder and CEO, Mihup-conversational AI company)
