Dutch airline KLM has partnered with DigitalGenius to help incorporate machine learning into its customer service, but humans are in no way getting replaced.
Robots remain a common theme in the ‘who will take your job?’ discussions accompanying much of the debate about the future of work. However, very few have looked at the complementary rather than the industrial, revolutionary side of things.
Monotonous, mundane tasks have forever come under attack from technology, with machines now operating across most manufacturing lines – though that doesn’t mean we should all fear change.
Take customer care, for example: it’s something that seems menial, but would struggle to fully work on automation. The banking industry embraced recorded contact before most industries, but automated phone calls have been the source of consumer frustration for decades now.
However, it’s hardly a perfect model, with disruption throughout the fintech food chain meaning any service not appealing to customers will quickly get overtaken.
Machine learning: Come fly with me
KLM has cottoned on to this and its new partnership with DigitalGenius is an attempt at evolution, as machine learning creeps deeper and deeper into our society.
DigitalGenius software has processed 60,000 questions and answers used by operators in recent years; ranking every word, sentence and phrase as number vectors that, when recalculated, provide potentially satisfying responses.
After a question is asked, these responses appear for the customer care agent to pick from, tailor and correct. This means the machine is doing the laborious work of finding the general resolution, but the agent does the rest.
“A personal approach is extremely important to KLM as this is what defines our social media service,” said Tjalling Smit, senior VP of digital at Air France KLM. “Applying AI, KLM can handle a greater volume of questions while still maintaining its personal approach and speed.”
With 235 customer service agents operating over 12 languages, complementary software such as this appeals to some. With every response, the AI improves, making it easier for KLM staff to keep customers on side.
Machine learning: Social connections
“Customers want to talk to companies in whatever way they want, and KLM did a great job enabling customer service over social channels,” said Mikhail Naumov, president and CSO at DigitalGenius, noting the opening of a Facebook Messenger tool as particularly significant for KLM.
“They got flooded by requests,” he said. “The spike in volume is quite expensive if you deal with it in just agents. You can only add more agents for so long. They wanted to innovate.
“So their data was all being logged, but not used. Now we can take historical logs and put them into a deep learning model which can understand language, despite phrasing, and help out.”
It’s actually quite clever when you think about it. Apps are on the way out, with conversational commerce the front runner as a replacement for them.
“We believe that we should be where our customers are,” said Smit in Forbes recently. “They spend hours daily on their mobile phones but not necessarily on our website. Many travel with us once or twice a year, so the likelihood that they have downloaded our app is limited.
Machine learning: A better fit
Customers spend a lot of time on messaging platforms instead, such as WeChat in China, for instance. That’s where KLM wanted to provide a service, needing AI to help keep the show on the road.
Naumov believes this is a different model and a better way of doing things, rather than chatbots, which he argues are too reliant on keywords.
“They can break,” he said. “This is because they are built using basic linguistics. They will work well in very defined areas – marketing or weather, for example – but when it comes to customer service, you can’t trust it.
“Anything can happen in customer service. Machines should take care of machine-level tasks. Say statistical logging, things that when humans do it, it’s so repetitive [that] they get bored.
“Humans should be employed to do human-level tasks, stuff that they do better than robots. KLM is essentially increasing the likelihood of happy customers, but also enhancing the process for their agents.”
Salesforce, according to Naumov, is one example of where this complementary, hybrid model is thriving. And there are many spaces, much like customer service, that could host the next stage of this process. Sales would work, he said, as there are plenty of repetitive conversations between humans.
“Law is a great area,” he added. “Creating documents over and over again, these are all things that have been done millions of times before. If you train a deep learning model here…”