Article content
Disruption, disruption, disruption.
It’s been only a couple of years that we hear this word almost everywhere and about potentially almost everything. This moment was supposed to come, at some point – the moment when we actually take a look at the meaning on a Cambridge dictionary. A disruption is “an interruption in the usual way that a system, process, or event works”. And it is true, the tech and industrial revolutions of this decade are indeed bringing dramatic changes to the way we do things: with augmented reality, with blockchain, with Artificial Intelligence.
But we’d like to suggest a further semantic development to this narrative. Instead of letting disruption be the dominant keyword of our times, we may want to take things a step further and talk about transformations.
Because that’s what we see happening around us – transformations. Processes don’t just terminate but evolve. There’s no irreversible or traumatic interruption but just change. And by looking at all the possibilities that new technologies bring with them, then our duty is to wonder how we can make these replace outdated practices, have an impact on production cycles, and increase the level of security of communications and transactions.
Beyond the hype. AI and customer satisfaction
In Estonia, we’ve already pulled Artificial Intelligence out of the hype cloud when we legalized self-driving cars and started developing AI-based delivery robots. Now, the Estonian company AlphaBlues is offering AI live chats and virtual assistants which will free humans from the boring job and let the machine learning keep your customer service satisfaction high.
Two years old, but already a series of successful case studies proving that the guys know to the bone what they’re doing. Indrek Vainu, the Harvard educated CEO of the company, showed us how AlphaBlues solutions work together and what kind of needs they are meant to address. The company’s flagship products are AlphaAI and AlphaChat, dealing with customer service, sales, and form filling. What makes them a unique presence on the market is the fact that a virtual assistant and a live chat are here integrated into a single, coherent solution.
“This enables to fully automate the customer’s journey into one flow – from website visits with a virtual assistant to a chat with a human agent. The depth at which we can do end-to-end chat automation is unparalleled. And the cool thing is that AlphaAI learns automatically by observing human agents answer questions in AlphaChat, and automatically trains itself”, Vainu explains.
And it’s here that we dive into the – often obscure – topic of machine learning. How does it work? Who is it going to learn from? What is going to learn? Interestingly enough, Roman mythology can help us better understand this process. Janus was the two-faced god of beginnings and ends, looking at the past and the future: the point in between, the passage, is what was leading people and things to a next step, but always looking at the two sides of history – or of the story. AlphaBlues solutions work in a similar fashion, by learning from questions and responses, from customers and service agents.
Training Artificial Intelligence becomes then a two-sided process, “where back-end operators simply use the system and the AI observes them in the background and gains experience so that the front-end customers could get even more accurate answers. It’s a very frictionless training system. We also included in our product a dedicated tool where AI trainers can label incoming user messages with one click, so to teach the bot what different messages mean and how to respond to them in the future”, Vainu says.
A cross-market revolution
The company has by now a list of successful case studies that includes LHV and Monese in the banking and finance sector, Lattelecom in telecommunications, and Tallink in marine transport. Currently, it seems logical that one of the most direct applications of AI-based solutions takes place in the customer service sector, as this involves the performance of an often repetitive work that could easily be carried out by a well-trained machine.
Two other reasons complete the picture: “Human labour is expensive in answering chats. And secondly, people are using chats more and more. There’s an increasing switch from calling and emailing companies to chatting with them, and the need to manage these chat volumes in an efficient and cost-effective manner is clear and out there. AI will automate many repetitive queries, but we’re moving beyond that. We’re building systems that are capable of talking with customers 24/7 and prioritize those conversations”, Vainu continues, “so that companies can grasp the true value in these interactions that otherwise, with AI and chatbots alone, would go lost.”
AlphaBlues is already expanding their market to other Central and Eastern European countries, filling a gap left by competitors focusing mostly on the widely spoken languages of the world, such as English and Spanish. Polish, Czech, and Latvian among others, fit well with a solution that was born in Estonia, a country that potentially faces similar difficulties in implementing bots referring to non-major “difficult” languages.
But the potential for expansion is not only a matter of market voids. AlphaBlues solutions represent a practical response to a practical necessity in a specific sector, addressing concrete needs of companies and partners willing to automate their customer services. What the company is doing calls into question the extent to which we’ll still rely on human labour for certain tasks. Can AI free humans from doing the boring job? “The tedious tasks that we can do in a few seconds can be completed by Artificial Intelligence. This frees up our time to do more meaningful work, tackle the tough problems that require creativity to reach a solution, and ultimately provide us with more satisfaction in our working environment”, Vainu states.
So it’s sorted, we’re safe: machines won’t take control of our governments and rule us all. For now, we can just make them work for us, and everything will be just fine – and more efficient.
