Why advances in natural language processing spell the end for so-called ‘user error’

Artificial intelligence is catching up with human intelligence. Take natural language processing (NLP). Now at the forefront of emerging consumer technologies, such as conversational AI, NLP has transformed the user experience.

Where we once had to think carefully about the right keywords to use when trying to get answers from technology, NLP enables us to ask questions of popular devices using the same language we would in everyday conversation.

Advances in machine learning in recent years have seen AI take on higher level human functions and AI-powered tools become an accepted part of our daily lives.

With the pandemic accelerating decades worth of digital transformation within the space of months, AI is fundamentally changing our relationship with technology and mainstream use cases for machine learning continue to grow exponentially.

NLP technology is just one of the areas that has come on leaps and bounds, but that doesn’t mean we should take it for granted.

Back in the early days of search engines, NLP technology was primitive. When users searched for a word or phrase in Google, it wasn’t uncommon for them to get three dreaded words in response – ‘no results found’.

Nothing was more frustrating for users than knowing the information you wanted was at your fingertips, but not having the magic words to access it. The best part was Google made us think this was all par for the course. It wasn’t so much a pain point but an accepted part of the online experience.

As users, we didn’t challenge the wisdom of supposed ‘user error’ or stop to wonder why the search engine hadn’t understood what we wanted. We simply told ourselves we’d not been precise enough to get the results we were after.

Over 20 years, we’ve become used to thinking like machines when we interact with technology. What a brilliant bit of cognitive conditioning by the tech giants!

Staple of engagement

This ‘think-like-a-machine’ mindset has underpinned human relationships with technology since the dawn of computers and coding but proliferated with the explosion of the World Wide Web at the turn of the century.

Many of us, for example, will be familiar with the feeling of frustration after selecting the wrong delivery address, or accidentally putting an item in our basket multiple times when shopping online.

Surely in 2020, the tools and software designed to make our lives easier should understand human behavior, our imperfect nature and our natural ways of communicating? Getting answers from technology, and completing the tasks associated with them, should ultimately be as easy as asking a question of another human being.

Thanks to the evolution of artificial intelligence and machine learning, we now have the ability to recognize and process natural language in online search. As technology better adapts to our needs as human beings, we’re also changing the way we interact with it to get things done.

According to Google, conversational search has led to an increase in personalized queries related to the needs of the individual. For instance, mobile searches beginning with “do I need” and “should I” have grown by more than 65 percent in recent years, while 70 percent of requests to Google Assistant are expressed in natural language.

From self-serve portals, to chatbots and virtual assistants, automated customer service is now a staple of the way we engage with brands and content online. But there remains an expectation-reality gap in this process.

If the questions we ask are difficult to understand, then we’re asked to clarify with a return question, often sparking an entirely ancillary conversation to pin-point exactly what it is the user wants.

The problem with current automated customer service tools stems from the fact that many providers use off-the-shelf chatbots or generalist NLP engines from providers like Amazon and Google.

A long way from keywords

The results are effectively glorified FAQ systems that lack the depth of understanding to handle complex, specific user queries in relation to business-critical services.

Automating broad, wide-ranging conversations between humans and an AI interface to a high degree of accuracy is still tricky. But we can narrow the fields in which NLP platforms, and the machine learning capabilities associated with them, are deployed.

By focusing on specific areas of knowledge or industries, it’s now possible to harness the power of proprietary NLP platforms that better fulfil user requests within these verticals, for example, using conversational AI. And, since NLP is context-driven, users can get the answers they’re looking for, even if their question isn’t spelled or worded correctly.

Fundamentally, it’s this capability to emulate truly human-like conversation that will take the user experience to the next level. But there’s another opportunity here for brands to think carefully not just about how their customers communicate, but the channels they use – an area in which AI is opening up new possibilities.

With rising app fatigue in an attention economy, forward-thinking businesses are looking to some of the most widely used apps on the planet, such as WhatsApp and Facebook Messenger, as the next frontier for user engagement.

Combining NLP with the most popular messaging apps and companies’ existing IT infrastructure – including order management and customer communication systems – has the ability to transform the entire user experience.

Lightning quick time to respond. Fully personalized and contextualized replies to queries. Intelligent suggestions that reflect what the user actually wants – not what you assume they want. It’s an exciting new world, and we have the tools to make it a reality right now.

From hyper-personification, to simpler ordering to enhanced customer service, AI can bring brands and their audiences closer together on the channels they use and love the most.

These are the kinds of user-focused experiences we must deliver in the era of AI. An era where technology adapts to us, not the other way around. And where machine learning can finally make the leap from servicing user needs, to anticipating what we want – even before we know ourselves.

Thankfully, we’ve come a long way from the days of keyword search. While AI still has some way to go in truly understanding human behavior, we no longer have to accept so-called ‘user error’ as we embrace a new, more complex relationship with technology.

Ed Hodges, CEO, HelloDone

Source link