This week, I had the pleasure of attending my first Gartner Symposium in Barcelona. A huge industry event, hosting around 7,500 CIOs and IT executives from around the world, that explores the technology and trends that will shape the future of IT and business into 2020 and beyond.
Interestingly, this year’s demographics saw the Nordics represented by around 20% of attendees – impressive for a region of only 27 million people collectively. Compare that to the UK and Ireland, who made up only 19% of attendees, but boast over double the population, at 63 million. Perhaps the reputation of the Nordic countries being amongst the most tech-savvy in the world is well earned?
Clear evidence of this eager embracing of digitization is apparent in the sheer number of enterprise deployments of conversational AI in the Nordic countries, which are easily a few years ahead of the rest of Europe. Most banks and insurers in Norway, as well as many government agencies, already have a virtual agent visible on the front page of their websites, offering customers and citizens a fast and efficient method of support to supplement (and often replace) traditional phone and email channels.
Of course, even in the Nordics, it’s still early days for the adoption of conversational AI across the entire enterprise space. We are only scratching the surface of what is possible to achieve with this game-changing technology, as the conference’s opening keynote presentation demonstrated.
Coining the term ‘Techquilibrium’, Valentin Sribar, Senior Vice President at Gartner, stressed that it was crucial for businesses to find the right balance between the ‘traditional’ and ‘digital’, which is in constant flux.
This balance between legacy systems and emerging technologies – and the symbiosis of vendors and customers – were recurring themes throughout the conference, along with a number of other key takeaways and trends related to conversational artificial intelligence that I am eager to share here.
Face your dilemmas and turn them into opportunities
‘The best defence is a good offence’ is a tried and tested adage that rings true when it comes to digitization. While building a business case is a common tactic when adopting new technology, it may not always be the right solution and ‘learning as you go’ can often be more rewarding in the long term. Working at the forefront of new technology like conversational AI can be challenging, as there are no manuals or clear instructions on what to do. And even though a vendor may strive to develop the very best technology possible to satisfy a need in the market, it’s crucial that a business (along with its decidedly human resources) can take charge in molding a virtual agent into an asset for their organization.
If you think of a virtual agent as a rocket ship, conversational AI is the engine that powers it. But without astronauts and mission control (i.e. humans) a successful launch just isn’t possible. All three pillars need to work together to guarantee a successful launch, and when it comes to artificial intelligence we mustn’t forget that humans play just as pivotal a role as algorithms and interfaces.
The only way to scale well is with the right technology
Today’s consumers are more demanding than ever. They expect the brands they interact with to be available at their convenience and to offer them instant access to solutions wherever they are.
In the Nordics, virtual agent interactions in the public sector peak when most human employees have already clocked off for the day. This means that without conversational AI, countless inquiries might go unanswered, and shows that if you give people the option to reach out when they have ample time to do so, they’ll gladly take it.
It’s impossible to scale up human support quickly without incurring enormous costs – only technology can increase a business’ capacity to handle large volumes of traffic.
AI is creepy when it comes across as ‘too human’
It may be a false endeavour to attempt to make your virtual agent come across as too human. Customers don’t like it and can easily see through it, and you end up bumping up against the uncanny valley effect.
People still expect artificial intelligence to be humanized, however it needs to be done correctly for them to accept. Studies have shown that customers prefer a business to be transparent and to show them clear signs that they are interacting with a machine. Many still want some form of human-like contact, such as a friendly tone, or a joke when they ask for one.
In contrast to what you may have seen some vendors claiming, we are not even close to humanoid robots being commonplace. What is interesting, however, is that customers are, for example, more likely to accept that their bank’s virtual agent is human-like when interacting through online chat or a voice assistant, but not if the same technology is present in physical form.
Augmenting humans
Much of the buzz surrounding AI in customer service has been about how the technology may end up replacing human employees. But is that really the goal? Of course not. The real power of artificial intelligence in business is its ability to augment humans and help enhance their already considerable skills set.
In customer service, it should never come down to choosing between a virtual agent and a human. Forward-thinking companies are today recognizing conversational AI as a powerful tool for collaboration. A virtual agent can assist in detecting sentiment and making suggestions on which answers to give. AI should be looked at as a key part of any customer interaction – whether it’s between a customer and a virtual agent, or with human support staff.
When Norway’s largest financial newspaper, Dagens Næringsliv, ran a headline around three years ago that read (and I’m paraphrasing the translation) “Bank robot replaces 40 people in customer service”, the initial hysteria eventually died down. The reported replacement didn’t end up happening (surprise, surprise!) and instead, the AI today works alongside the employees at the bank mentioned in the article, helping them to deliver exemplary customer service on a much larger scale.
Finding the 5 percent
“95% of bots in the market are s***” was one of many bold statements made during one of the presentations that I attended over the course of last week’s conference. And it is the one that resonated most strongly with me.
Are all bots created equal? Ask any company that has struggled to take a chatbot project from POC to production, and they will resoundingly tell you no. It’s these bad bots that are fouling the name of the wider conversational AI market.
In today’s chaotic marketplace, finding the 5% of vendors that can deliver a scalable virtual agent that has robust language understanding capabilities, watertight security and privacy compliance and enterprise features might feel like looking for a needle in a haystack. The struggle is real, but, ultimately, the potential for conversational AI to help companies attain a ‘techquilibrium’ is a journey worth seeing through to the end.