Computers and robots that can talk have long been a staple of science fiction – from Lost In Space and Star Trek, to 2001: A Space Odyssey and Blade Runner. The father of conversational artificial intelligence (AI) – Alan Turing – famously posed the question, ‘Can machines think?’ in his 1950 paper ‘Computing Machinery And Intelligence’ and devised what is generally referred to as the Turing Test, an ‘imitation game’ where a human interrogator attempts to determine which of two responders is a human and which is a computer. Even advanced AI language models like GPT3 and BERT still can’t convincingly pass the Turing Test on a wide range of topics – at least not yet.
Natural Language Processing
The ongoing and rapid evolution of Natural Language Processing (NLP) means that day is getting ever closer. NLP – generally powered by machine learning – refers to technologies designed to interpret the underlying intention behind what a person is asking, regardless of how they phrase it, and respond appropriately. The leading conversational AI platforms can now conduct sophisticated multi-turn-based conversations, listen for and remember key identifiers (like phone numbers, post codes, account numbers), handle interruptions smoothly, remember context from previous conversations and perform actions on request.
The improvement in voice technologies
Voice technologies – specifically Speech to Text (STT) – have also improved dramatically in their accuracy over the last few years, with Google, Microsoft and Amazon leading the way. Probably as a direct consequence, the adoption of voice assistants – and other voice enabled devices – has been on the rise during the same timeframe. The pandemic and its lockdowns also contributed by accelerating the use of voice interaction, converting infrequent users into routine ones.
The same conversational AI technologies behind those sophisticated voice assistants can also be applied to the classic voice channel – the telephone – enabling a genuine conversation between a customer and a virtual customer services agent, instead of the ‘press one for this, press two for that’ endless menus of options that tend to confound customers. Authenticating a customer over the phone can be further simplified by use of voice biometrics – the identification of a customer based on their voice – which can deliver a much improved caller experience.
Multi-channel conversations
Using an enterprise level platform with a range of deployment channels means you can expose the same conversational capabilities through all your communication channels, providing a single consistent way for your customers to interact with you. Those channels can be anything – Facebook Messenger, WhatsApp, SMS, Twitter, Google Assistant, telephone, email, Teams and not forgetting the classic web chatbot – and the principle is to implement once and deploy everywhere.
Robotic Process Automation
If conversational AI is the brains of your virtual agent, API integrations and Robotic Process Automation tooling like UiPath or Blue Prism are the hands. These empower your virtual agents with the ability to authenticate a customer, provide information and action requests – for example, confirming a customer’s identity, allowing them to change their contact details or taking a payment.
Augmenting your contact centres with NLP powered conversational AI means that the most common asks can be handled 24-7 by a limitless virtual army of AI agents while human colleagues can give more time to those customers that really need them. All this reduces call waiting times – and caller frustration – as calls for the most common customer needs can be dealt with purely by the conversational AI agents.
The tech has finally caught up with our sci-fi dreams of a Star Trek computer that we can talk to, can give us information and can actually do things for us. With the increased level of customer service interactions that were a feature of the pandemic and now the evolving cost of living crisis, there has never been a better time to give your customers and your colleagues some virtual help.
“We can only see a short distance ahead, but we can see plenty there that needs to be done.” – Alan Turing
Want to learn more? If you would like to learn more, please reach out to KoKo Visser, Partner and CAI Expert at Valcon, at [email protected] or call +31623609586 today.
