From chatbots to programmatic marketing, artificial intelligence is booming in marketing. But how do we define ‘AI’? And are there different levels and applications of machine learning technology? Ryan Lester, Director, Customer Engagement Technologies, LogMeIn explains the different types of AI capabilities that exist when it comes enabling customer service.
Today the term AI encapsulates an increasingly large umbrella of technologies. Many companies are jumping on the AI bandwagon, but not all AI is created equal. For example, the type of AI which powers your portable home assistant may be very different to the AI which drives an autonomous vehicle. Without really digging into the bits and bytes, it can be hard to understand the difference.
Recent research indicates that 88% of companies have either invested or are planning to invest in AI technologies for customer experience, but which one is the right one? Currently, there are three main categories of AI that most CX technologies fall into — scripted, cognitive and conversational. Let’s dig into each a bit.
Scripted AI
Scripted AI is the technology that often powers most of the traditional AI we are familiar with today. IVR systems and even Alexa run on this type of AI. Scripted AI relies on preloaded comments triggered by certain words and phrases. Think “Alexa, play U2 on Amazon Music”. Alexa doesn’t actually understand the sentence being asked, it’s action is triggered by the words being used. Play. U2. Amazon Music. However, if a customer diverts from this script, Alexa likely would not understand and couldn’t perform the action. While scripted AI does provide a short-term solution, it can’t scale and will be unlikely to meet customer’s growing expectations. Since these bots don’t really understand the intention behind the words, but rather the words themselves, it would be impossible for a company to train the system to understand all variations of all possible questions they could get asked.
Cognitive AI
Cognitive AI lives at the other end of the AI spectrum. While only at the start of its development – the possibilities of what can be done with cognitive AI are seemingly limitless. However actual attempts of real world implementation have been an entirely different story. Cognitive AI uses complex machine learning algorithms and Natural Language Processing (NLP) to learn on its own and have human-like conversations. While these technologies have great promise, they come with associated risks. Being able to learn on its own and take actions based on what it knows without human intervention can spell disaster. Businesses need to have complete control over the outputs, or the system could end up doing something you wish it hadn’t. For example, a bot could learn that free products make customers happy and therefore start giving all products away for free. Anything is possible for cognitive AI solutions, and that’s not always a good thing.
Conversational AI
Conversational AI is powered by structured data and leverages natural language processing that allows the bot to understand the intent behind a question – not just the words being used. This means that no matter how a question is phrased, the bot understands and will provide the same answer. Conversational AI falls in between pre-written scripts and completely unstructured machine learning to offer a the ability to learn from limited, structured data while a human supervises to ensure nothing goes awry.
With Gartner predicting that over 50 per cent of medium to large enterprises will have deployed AI-powered chatbots by 2020, companies need to assess what variation of AI will work for them. Conversational AI provides a solution that is just right, combining the best of both worlds to form the ideal customer experience solution. Cognitive AI is too hard to predict and resource gobbling, while scripted chatbots simply can’t provide the high-quality experiences that today’s customers demand. Conversational chatbots deliver a mixture of human supervised machine learning and NLP that provides a high levels of customer facing acumen and customisation needed to satisfy today’s customer experience connoisseurs.
By Ryan Lester
Director, Customer Engagement Technologies
LogMeIn