Artificial Intelligence (AI) has made the leap from hype and buzz to become a useful business tool. But the technology is new enough that enterprises can have a tough time figuring out when and how to use it. Great advice on how to do that comes from a recent Gartner study, “Where You Should Use Artificial Intelligence — and Why.” It covers a wide variety of topics, but some of the more interesting recommendations are related to mobile, as I’ll detail in the rest of this blog.
Here’s some advice for determining when to use artificial intelligence in mobile apps you’re building.
First, a little background. The report notes that AI is increasingly becoming top-of-mind for many companies, reporting that “Interest in artificial intelligence (AI) has accelerated markedly; the volume of client conversations on the topic increased 200% from 2015 to 2016.”
As for advice about how and when to use AI, the report recommends, “Employ AI to solve challenges in which you lack the resources or corporate worker base to succeed. Identify repeatable tasks where the need is repetitive but the outcomes vary, for which AI is particularly useful.”
What does that have to do with when to use artificial intelligence in mobile apps? Plenty. The report did a survey to determine what kind of application categories enterprises say will gain the most from integration with AI. The top one, cited by 34% of respondents, was customer engagement applications. That’s exactly what mobile apps are designed for. In addition, 23% of respondents said they would integrate AI with digital marketing platforms – and mobile apps often play a key role in them.
The report recommends that AI be used to build virtual private assistants, which are one of the hotspots for mobile these days. It says that AI and machine learning should be used with these assistants, “in automating the personalization of marketing and merchandising messages.”
It also had recommendations for how to best deploy AI, and again, it has some good advice for anyone building mobile apps that interact with customers. It says enterprises should “Identify narrow, routine customer interactions,” because these will be easiest to automate with IT. It then adds that virtual customer assistants need to be tested to make sure they’re stable and actually help customers, before fully deploying them. And it warns, “Gartner predicts that poor-quality implementations could significantly increase customer dissatisfaction.”