With all the hype surrounding artificial intelligence (AI) these days, it's easy to be overwhelmed. Will AI replace my job or just help me get a better one? Can it really make sales reps smarter, customer service agents faster, and customers more delighted, or will it dehumanize customer interactions and drive us to Stepford? Whatever side of the argument you're on, there's still a lot of confusion about what AI actually is and does. To put things in a more grounded context, I like to think about AI like I do about another non-human that significantly impacted my life: my dog.
AI without training is like a new puppy without the cuteness: pretty useless. But like that puppy, with training it can learn to do a lot on its own. In analyzing hundreds of AI projects over the past two years, I've found that we've come a long way very quickly. From simple predictions and recommendations to complex image recognition, AI is rapidly becoming a part of everyday work and personal life.
Roll up the newspaper again and again.
Anyone who's gotten a puppy knows that the first months can be a real challenge. How many times does he need to be told not to chew shoes before he gets it? Bribery goes a long way, but so does ongoing positive and negative reinforcement. Like puppies, AI is trainable, but it only learns from the belly rubs and rolled-up newspapers it gets. And it needs a lot of them. The great news is that customer service data is great for AI training: it's fairly structured, has common themes, and with natural language processing getting better and better, it can rely on tons of new data on a regular basis to get smarter.
You can skip the crate training.
As AI applications have developed, there are a lot of prepackaged apps that have already been trained. Chatbots are great examples. Although they're not the ideal customer service agents for all interactions, many already have a broad enough vocabulary and process knowledge that they can hit the ground running pretty quickly. Prepackaged capabilities, such as prediction engines and self-learning knowledge bases, also enable you to skip much of the basic training and focus your puppy's energy on objectives specific to your business.
It's not them; it's you.
A professional dog trainer once told me all my dog's "problems" were about me, not him (that was the first and last training session we had with her). Effectively leveraging AI requires effective training, both for the technology and the humans that train it and interact with it. The good news is that there are a wealth of resources to get you up to speed on AI (for instance, one practitioner told me he learned everything he needed to know about Salesforce Einstein from a weekend of YouTube videos).
With the proper training, a dog can be a loving companion, a source of comfort, and a great source of exercise. While AI might never meet those objectives, when trained properly it can probably get close, and you won't have to replace the rug.
Rebecca Wettemann is vice president of Nucleus Research.