Will AI be the new face of customer services?

AI (Artificial Intelligence) is rapidly becoming a technology that changes how a business works and interacts with its customers and users. It’s increasingly becoming a viable solution to wide-ranging problems, including customer services.

At the their most recent IO event, Google introduced Duplex — an AI system for accomplishing real-world tasks over the phone. In an impressive demonstration of an AI assistant making a phone call to a hair salon, Google introduced what may become the new norm — interactions where customers can’t tell if they are talking to a human or computer. The demo was a stark example of how far AI has come in terms of understanding content and the context of customer requests and questions.

Imagine Google Duplex as a tool for customer service. It can distinguish tone of voice and context — it’s able to intelligently ask relevant questions and drill down into issues. There’s no reason it can’t be used to drill down into problems and offer solutions to simple issues without any human intervention.

While the unstaged Duplex demo was impressive in its use of voice, this isn’t the first instance we have seen of this kind of technology — people are using AI assistants today to solve similar types of problems. Microsoft happens to be one of these organisations, and they’re using AI to deal with customer support issues today.

Microsoft offers online support for all of their products, so imagine a customer struggling to add a printer to Windows 10. As the user browses for support, a chat window appears with what seems to be an engineer offering support. The conversation opens with brief pleasantries —with the engineer asking about the issue. The customer explains the problem, and the engineer starts to give simple instructions to try help them out.

This whole conversation is being handled by AI.

At no point is the customer made aware, nor feels like they’re talking to an AI, but they are. As the conversation progresses the AI is unfortunately unable to help and suggests connecting the customer to a colleague, in this case a real human.

Instead of the conversation ending here, this is where the real magic happens. The AI observes the conversation between the two humans — and learns what the human engineer does to resolve the issue. Once resolved, the AI understands the steps towards resolution for the next time this kind of support request is intercepted.

Over time we should expect that all 1st line support would be handled by AI, allowing support teams to work on more complex issues and projects. With AI shifting from text to voice, it leads me to wonder if we’ll end up missing out on incidental human interactions.

It’s worth considering the impact automation has already had on people. Take ATMs as a long-lived example- developed in the 1960s and ubiquitous by the 1980s , ATMs have replaced a vast majority of human tellers. They’re common, they’re the norm, but even in 2018 banks still have tellers. There’s still a proportion of the population that want to talk to a “trained professional” over a machine.

People still matter. Even though there are plenty of studies showing AI outpacing trained professionals in a huge number of fields, it doesn’t mean that people always want to interact with machines instead of humans. People are emotional.

One San Francisco restaurant has taken people out of the equation completely, replacing them with iPads. After placing your order, your food comes out of a dumbwaiter, with no runners or human contact required. But it’s not for everyone — some people like to be welcomed, served and attended to.

We see similar things in services like Uber or Lyft. Driverless cars are a growing reality and car services will push in that direction because it’s cheaper for them to do so. It doesn’t mean all consumers will enjoy a driver-less ride. People often like the experience of talking to drivers- they get more from the experience than just transportation.

Technology marches relentlessly onward, automation and AI will come, but some things remain fundamental, and people-to-people communication will continue to be one of them. Just because the tech is available, doesn’t mean it’s always going to be the best option. AI and automation opens a vast number of ethical dilemmas, and the value of human communication is just one of them.

We might not be in a place today where AI offers the catharsis of yelling at support — perhaps it would even cause more frustration if caught out, but I believe that over time the march of AI (or even the emergence of robots) has a good chance of changing that fact.

Until then, we’re still going to keep a few humans around.