The AI market is growing at an exceptional rate and is set to reach $16.06 billion by 2022. IBM has been at the heart of AI for several years and with its AI platform, Watson, the tech giant is now enabling developers to tap into its power, unlocking a world of potential for brands. We spoke to our Co-Founder, James Cannings, to get his views on the cutting-edge platform and how it could revolutionise the customer journey.
What is IBM Watson?
Watson is the outcome of IBM’s ongoing work in the field of artificial intelligence (AI). AI has become a buzz phrase and there’s a lot of hype around it, but IBM has genuine credentials going back to its Deep Blue project which beat chess world champion Gary Kasparov.
IBM has always been at the cutting edge of AI and Watson, named after the company’s founder Thomas J Watson, is the next step. Its natural language understanding and machine learning abilities allowed it to clock up another pop culture success when in 2011, it beat the two most successful champions of US gameshow Jeopardy. It learned to play and win the game by watching multiple series of the show, and participated without being connected to the internet.
What does it do?
Basically, Watson offers brands the opportunity to understand, analyse and interpret their data at an unprecedented rate. Because it operates through constant feedback, Watson represents genuine AI and machine learning - it’s not just number crunching. For instance, the system promises to understand the content and context of questions and how to answer them.
It offers natural language understanding of speech and analysis across millions of documents, such as user manuals. This allows users to ask questions in natural language. As well as responding appropriately, Watson can learn to understand the tone and personality, so if somebody is getting angry or frustrated, it can hand them over to a live operative, for example. It understands intent, even if it has never come across a specific sentence before. It can also recognise objects, using visual recognition to understand image based content.
In the past year, IBM has been pushing a set of API tools for developers to help them tap into the power of Watson and open up a broad spectrum of opportunities for brands to develop innovative marketing and ecommerce solutions and cost saving process tools.
Why is it exciting?
One of the most exciting aspects of Watson is that it allows brands to put the consumer at the heart of things, rather than expecting them to bend to technology. It’s far more than just a chatbot that you bolt into your website, although you can easily build one if that’s what your customers need.
Brands often struggle to deal with the customer journey across multiple devices although User Experience teams have been focusing hard on that over the last few years. Data has always shown us that people tend to browse on mobile devices and complete their purchase on a laptop, for instance. With the rise of voice enabled devices, such as Amazon’s Alexa, Google Voice or mobile phones, that situation will get more complex still. Alexa might be able to tell you where to find the cheapest phone deals, but if you want to buy it, the journey can come to an abrupt end. It’s a telling move that the new Alexa will launch with a screen which may start to help us break down these barriers.
Although Watson comes as a series of APIs, these are very open compared with existing chatbots which tend to be quite productised. They are basically add-on widgets, which may help cut costs, but don’t necessarily improve the customer journey.
With Watson, you can feed in any content and get it to learn anything. This allows you to build a whole web experience around the interaction with Watson and think about how you can make the customer journey faster and easier. For UX people it’s great.
Another plus is that the business model of Watson lends itself to being experimental. You don’t have to spend millions on a licence, so can get up and running quickly and cheaply. You can prototype for free.
Where will it be used?
One of the most common uses is for chatbots or ‘virtual agents’. When people think of virtual agents, they think of a box popping up at the edge of the screen, replacing a real person and cutting head count. However, brands need to consider how they can change the whole user experience.
You could build a virtual agent to walk you through an ecommerce process for instance. Diageo did this with Whisky Matcher, which operated on Facebook Messenger and recommended suitable whiskies based on a consumer’s flavour preferences.
Virtual agents don’t have to be like the bots we see currently. The experience might be less obvious than chatting – it could be drop down lists for instance, where simple choices start to flow through, powered by AI and not a preset navigational structure.
You don’t necessarily need to think of a conversation, but how a store is structured, for example. Sites have a lot of content and there’s traditional structure and taxonomy. However, a Watson virtual agent lets you navigate the journey in a more intelligent way.
Because Watson can pick up tone and analyse when people are getting frustrated, there is a real opportunity in areas like customer service, delivering a more personalised approach, especially knowing when to bring in a live operative.
How will it improve the customer experience?
Brands can step away from the device and look at how they can make the customer journey easier. It’s an opportunity to work with marketing teams to see what they want to do for the consumer. A virtual agent might be a series of prompts on a journey for example, and it’s learning every time. By carrying out a traditional customer journey mapping exercise, you can start with something simple and gradually move on to more sophisticated applications.
The sky’s the limit with natural language understanding and with voice search looking like the next big thing, the challenge, from the customer point of view, is to help the journey flow easily between devices. As well as process automation, Watson has an exciting ability to connect AI and voice across devices and allow customers to flow more seamlessly between voice or chatbot interactions and more traditional interfaces.