Virtual assistants become smarter

The hybrid work format has speeded up the use of intelligent virtual assistants, for customer-facing roles as well as internal enterprise management functions

Virtual assistants are becoming more and more realistic. Even as virtual meetings have replaced in-person meetings in recent months, the rise of virtual assistants has been even more remarkable. Virtual assistants, or digital assistants, have been around for years. In the new version, they are smarter, intuitive and far more capable.

The Tingwu AI developed by tech giant Alibaba can identify and distinguish between 10 different voices in a meeting. Tingwu can transcribe conversations, summarise content and correct its mistakes as it goes along. It can understand English and several Chinese dialects.

IBM Watson digital assistant has also been updated with new features, especially for companies that have to deal with the peak rush of customer queries. Call-centres and websites are often unable to match customer questions in peak sales periods. After answering queries on websites, the chatbot automatically hands over to a human agent with relevant background, where required.

Government app Umang, which delivers information and citizens’ services, has also begun to deploy virtual assistants to deliver support, including use of voice-based commands which can help citizens who may not be comfortable using digital technology.

Maruti Suzuki India has deployed a virtual assistant on its application for customers to get help during breakdowns and queries for troubleshooting. In many markets, Skoda has integrated virtual assistants with support across languages.

The hybrid work format has accelerated the use of such intelligent virtual assistants (IVAs) for not just customer-facing roles but for internal enterprise management functions. “An increasing number of people are working remotely and are highly dependent on virtual assistants to perform their tasks. Therefore, it offers enormous opportunities for market growth,” says a report by GrandView Research. “IVAs are efficient and perform various functions for enterprises across retail, IT, banking, automotive, and others. Additionally, they deliver better conversational responses, which enables improving the customer experience. The IVA vendors are actively working on accommodating regional languages and provide customised virtual assistants so that customers from different geographies can easily communicate.”

The IVA market is about $6 billion and is expected to grow at a CAGR of 28 per cent till 2028, GrandView Research has estimated.

Chatbots and virtual assistants are also powered by conversational AI. This feature allows the IVAs to respond to queries in a human-like manner. Behind the effort is a series of activities which enable the response. A request with an IVA is processed to understand the query. A human-like answer is then formulated while identifying the topic. Such efforts are more efficient than mere chatbots, which offer pre-decided answers.

There are also rising concerns about the rules governing voice-based commands and conversations with bots. In every conversation, the bot has to understand and match the voice with a person. This amounts to recording of the biometric information of a person. The identifiable voice of a person is as important as fingerprints or a retina scan. As voice-based instructions rise, especially in markets where privacy and personal data rules are not strong enough, the scope for misuse and abuse remains high.

Without realising it, consumers can reveal banking details or access codes of their devices or personal health information while using voice commands. As many more millions join the digital mainstream, the need for managing IVAs, especially those using voice, becomes important. The European Union’s General Data Protection Regulation covers voice biometrics, but much depends on the consent of users.

As companies enhance the use of IVAs and voice-based features, the regulations will have to get smarter as well.