Artificial Intelligence (AI) is a frequently discussed topic. Just about every magazine has written about the advancements, opportunities and, yes, even fears this technology represents. The industry has made tremendous strides in developing software that works, and there are plenty of examples of practical solutions ranging from speech recognition to autonomous driving.
Yet, when we look at the enterprise-level adoption of AI, the use cases are much less frequent. In fact, Gartner estimates that only 6 percent of organizations use AI in production today. One reason for the slow adoption is that there is still some level of doubt about the technology. Can it perform as well as humans? Another reason is that there is a sometimes healthy, and sometimes unhealthy, level of protectionism going on. Is it going to take away our jobs? I also suspect that the frequent talk about the need for data scientists might be intimidating the knowledge workers without advanced math or science degrees.
To address these concerns, enterprise software providers need to be very smart about how and where they employ AI. The approach has to be very different than what works in the consumer space. What’s acceptable inaccuracy when asking Amazon’s Echo a question would not fly in the enterprise. In the consumer world, it is part of the fun to find new uses for Amazon’s Alexa and to see what she can do (Echo is a she, right?). But in the enterprise, AI has to be employed with a specific purpose in mind, addressing a clear pain point.
How do we find such a pain point? As Nick Patience from 451 Research says: “Anything that has been encoded in rules is therefore an important business process and worth looking at as a potential use case for AI.” Specifically, AI can be impactful when applied to a repeatable, high value business process. AI can have a huge impact when applied to the points of greatest uncertainty — where a decision is required that today relies on experience, intuition or business rules.
Let’s take the field service delivery process as an example, specifically the scheduling of field engineers. This is a very important and very repeatable business process. There are two greatest points of uncertainty here — the drive time and the time-to-repair. In order to schedule a field engineer to do a job, you need to know how long it will take to get to location and finish the job. You get one of them wrong and you have an unhappy customer — and your entire future schedule is impacted!
The drive time estimate is provided by Google Maps, which also provides routing advice. That’s covered. The time-to-repair is tricky as it depends on the engineer, his or her skill, the type of job, the type of product, the location and many other factors. Until now, the estimate has been made manually by dispatchers who relied on their experience … and sometimes just guessed the mean time-to-repair for each job. Estimating the time-to-repair is a major point of uncertainty in the scheduling process.
So, how can ServiceMax help?
All of that is changing with the new ServiceMax Autumn ’17 release. In this release, we pragmatically apply AI technology to solve this business problem. Introducing predicted time-to-repair, ServiceMax now uses AI to provide the dispatchers with a highly accurate estimate of the repair time for each particular job. The more data there is, the more precise the prediction. And unlike a rules-based approach, the AI system learns as the technicians improve their skills and the prediction remains accurate.
By embracing AI, ServiceMax addresses one of the key points of uncertainty in the scheduling and dispatching business process. This is a very tangible, pragmatic use case for AI in field service management and it doesn’t require any data scientists. Sure, there are other decision steps in the scheduling process that can benefit from AI and we have plenty of plans related to parts planning, skills matching, pricing, etc.
And of course, there are many other innovations introduced in Autumn ’17, including the new capability to model job dependency, as well as the app-to-app communication API for the mobile ServiceMax Field Service App.
To learn more about the ServiceMax Autumn ’17 release, check out the webinar.