Artificial intelligence and machine learning has created hundreds of new opportunities to optimize business.
But what’s the easiest way to get started? What’s the “low-hanging fruit” — the applications for artificial intelligence that involve the biggest benefits for the lowest costs in the least time?
It’s typically not about expensive data scientists trawling through data lakes full of terabytes of unstructured information. Instead, the easiest opportunities are likely to involve automating existing business activities. Gartner, for example, believes that this year, half a billion users will save two hours a day thanks to AI-powered tools — that’s almost half a century of extra productivity!
Here are the three main criteria to look for to make your first AI-powered applications easy, fast, and high-value:
Easy: Use your existing, high-quality data. Machine learning algorithms work best with large quantities of high-quality data. The easiest opportunity is to use what you already have! Look for AI opportunities based on high-volume data that’s already been subjected to quality controls, such as finance, manufacturing, or logistics, in your existing business applications or enterprise data warehouse.
Fast: Integration with existing business process. Artificial intelligence holds out some amazing new opportunities for business transformation, and doing things that were previously impossible. But fastest way to get started is with AI that can be seamlessly implemented as part of existing processes. This means you should look for applications that can integrate into your existing business systems with a minimum of disruption to procedures, training, bonuses, etc.
High Value: Clearly-defined, repetitive decisions. Automation brings the most value when it’s applied to narrow, repetitive business decisions that are made thousands of times a day. The inputs can be complex, with many thousands of variables to consider in real time, but the output should be simple-to-understand recommendations. Look around the organization for boring, repetitive activities that up until now have been too complex to automate, such as: Is this incoming call more likely to be about a new subscription, a billing problem, or technical support? What bank payments are most likely to correspond to this invoice? What promotion should we offer to this customer at the till? What resumes should we shortlist for this open position? What prospect should I call next?
The bottom line: turn first to your existing application vendors. For all the reasons above, the easiest way to get started with artificial intelligence in your organization is almost certainly to work with your existing strategic business application vendors, who have the domain expertise and systems knowledge to make the application of machine learning fast and easy.
These initial applications might not be the most business transforming uses of artificial intelligence, but when it comes to machine learning in the enterprise, it’s time to get excited about the boring stuff!