For field service organizations, predictive maintenance is the future, a fact that was highlighted in a recent report from Persistence Market Research. The firm anticipates the predictive maintenance market will reach $34.1 billion by 2030 – comparatively, last year, the firm said market revenue stood at $8.6 billion.
Technology adoption is a major enabler of this growth – the Internet of Things (IoT), big data and analytics, and other solutions have been deployed to meet the need for more operational efficiency as customer demand increases and service deadlines tighten.
Some of these technologies have been in place for quite some time. Even before the term “Internet of Things” was coined, service technicians were leveraging data from connected devices to help improve service responsiveness and visibility. Edge computing in the manufacturing sector continues to improve as processors get more powerful, and cloud-based solutions have helped companies extend maintenance and service data across multiple locations. The Persistence report also mentions predictive analytics platforms, which have integrated predictive maintenance and real-time analytics to help enable this type of highly responsive service.
AI Adds a New Layer
What is new is the rapid evolution and adoption of artificial intelligence and machine learning in the manufacturing and service sectors, which can help further improve predictive capabilities through more granular analysis.
AI algorithms may also help technicians (or customers) on site with troubleshooting, diagnosis, and repair activities. An AI system trained on years of service data on a specific type of equipment (or a specific piece of equipment in use by a customer) can help technicians leverage existing knowledge much faster and take guesswork out of the job. This wouldn’t replace technician expertise but rather compliments it by narrowing down the scope of the problem and helping the technician get to the right solution faster.
While the promise of predictive service is immense, the report does also point to some of the challenges posed by predictive maintenance. Some of the logistical and cultural problems to overcome before we can fully realize the potential of predictive are:
Regulations: As we have seen with recent controversies around the ChatGPT AI platform and plagiarism, the data needed to feed these predictive maintenance systems may be covered by data privacy and security regulations. As a service provider, how much access to machine data at a customer site can you expect? If a manufacturer sets up a predictive maintenance contract with a customer, who owns that data? How can you be sure it won't be compromised in a data breach, and who is liable if it is? Is any consumer data exposed in these systems (as it may be, for example, with point of sales systems at a retail store)? Privacy/security issues are going to vary by industry, but they have to be ironed out for these predictive systems to work as advertised.
Data quality and cost. Predictive analytics require high-quality, uniform data, which can be a tall order for companies running legacy platforms. The upfront work to scrub and categorize data and implement the right tools can be time consuming and costly, a big hurdle for small and medium-sized customers.
Staffing. If you think field service has a problem with staffing shortages, you should take a look at the IT industry. For these systems to work, most companies will need someone on staff with a background in data science and AI, or they will need to work with a software vendor or third-party integrator that can support them.
Customer resistance. This somewhat relates to the data privacy issue, in that connected equipment at a customer site is going to look like one more node on the client network that could potentially be vulnerable to a cyberattack. Or the customer may just not want a vendor to have 24/7 access to their equipment. If the customer is in the retail, legal, defense, or medical sector, they may be prohibited from granting that type of access. In addition, shifting from a break/fix mindset to paying for predictive maintenance can be a challenge both from a change management perspective, and a budgeting perspective. Field service organizations will need to be ambassadors for this approach.
The upside remains significant: reducing downtime for customers and avoiding expensive truck rolls for the service organization, longer equipment life, and new types of service products that can provide reliable revenue streams for manufacturers and service companies. Further downstream, there could also be benefits related to improved visibility of repair/aftermarket parts demand.
For service organizations that want to restructure operations around predictive, the first steps have to be analyzing both customers' and internal needs to figure out where and how this approach can provide the most value. From there, companies will need to determine how open their customers are to that business model and address any regulatory hurdles. Where you plan for an AI component, you must be sure you’ve done the necessary work to organize the data required.
This is all before you select a partner and deploy any technology! Predictive service represents a big change for a lot of companies culturally, economically, and technologically – but there are plenty of good examples of companies that have successfully made that shift. I have written before about several of them – Husky, Makino, and TKElevator come to mind.
If your company has transitioned, or is transitioning, to predictive, I would love to hear about your experience. You can email me here.