Type above and press Enter to search. Press Esc to cancel.

April 12, 2019 | 4 Mins Read

The Many Brains Behind AI in Service

April 12, 2019 | 4 Mins Read

The Many Brains Behind AI in Service


By Tom Paquin

Last year, I wrote an article in which I compared AI in video games—a phenomenon that’s been around for over 40 years—with AI in business applications—a comparatively new development. In it, I used the example of a two-brained approach to AI development. One “brain” manages an operational task, while the other manages a customer-focused task. I argued that this, ultimately, is what AI in service should look like.

In exploring the most recent developments in the Service Management space as they work towards more AI-powered systems, we can begin to see this come into fuller view. In doing so, it’s become clear that AI, as it will exist in the future of Service, will consist on significantly more than two interlocking brains.

To understand how this multi-tiered AI model will work, we need to understand how AI embeds itself into service management software today.

Let’s start by defining AI. AI is less like a piece of software itself, and more like a complex, open-ended programming language. The main difference between AI-powered functions and traditional programming is whereas traditional programming is designed around generating specific output based on specific input, AI programming is specifically designed to create unique output, based on multiple factors, in response to input. As more data sources are made available to the AI system, the output my change in ways that the programmer would not have initially considered. This allows AI-powered applications to “Learn”, grow, and, ultimately allows for programs and systems to provide much more dynamic output than a programmer could build into a system on their own.

So, what does that look like today? In service, it looks like optimization systems built into certain applications. For instance, your routing system may improve over time as it learns routes, traffic patterns, and times to complete jobs in order to more effectively automate scheduling. Or your Customer Experience system will improve to the point that it’s able to recommend customer service solutions based on what, previously, has netted the highest NPS.

These things all exist today, in one way or another, and each represent a single, little, AI brain. This is phase 1 of AI penetration into Field Service, and is already helping organizations streamline operations and eliminate the needs for backoffice staff to handle simple jobs. Phase 1 will conclude when these systems of automation reach a point of reasonable maturity, and, more importantly, organizations start looking at ways that central operations will automate themselves.

Phase 2 will of course be that centralized system; a brain that automates how all of these individual systems work together. Here’s an example: A sensor-enabled device indicates that asset has been running to the point that it’s time to schedule maintenance. The central AI system takes that AI brain’s analysis, and uses it to schedule an appointment. The routing system’s brain then dispatches a technician, offering a concise service window based on previous performance. It provides the customer with updates. Upon job completion, it checks with the inventory system, and invoices appropriately. All repetitive tasks are automated, allowing staff to focus on complex engagements, the act of service, and customer engagement. We’re not there, yet, but this is the future. A central brain working with dozens of AI systems embedded in dozens of apps, playing into each other, developing a common language, and executing on service before a human could input their name.

Sure, this is the future, but it behooves organizations to begin looking at these things today. To that end, there are three major things that organizations should be focused on:

  • Evaluate what AI systems you have today. Since those sub-brains that will inform the overall AI brain that runs operations in your business exist in so many applications today, it’s important to take stock of those platforms. Are the AI systems in place performing at an optimal rate?
  • Evaluate the integrity of your connected devices. If you work in an asset-intensive business and you’re not collecting and synthesizing asset data, then you’re way behind the curve. That’s not the only connected element of a service organization, of course. How about vans, warehouse inventory, and, importantly, your mobile workers.
  • Evaluate the overall connectivity of system processes. For the connected AI of the future to manage all of your systems efficiently, your systems need to connect to one another. Does your routing system speak to your order management system? Does your inventory management speak to your fleet management? Interconnectivity is no longer a nice-to-have. The AI-powered future calls for a unified system.

With these things in mind, you will be on the right path to set yourself up for AI success.