Predictive maintenance is emerging as a critical strategy for fostering more reliable, more efficient, and safer transportation networks. By using advanced technologies such as sensors, artificial intelligence (AI), and advanced analytics, road owners and operators can more effectively foresee and mitigate issues. This approach yields significant cost reductions, preemptively addresses safety hazards, and boosts availability of assets.
On June 28, McKinsey’s Global Infrastructure Initiative (GII) hosted a virtual roundtable for senior road developers and operators and public-sector leaders across Europe on the theme of harnessing the power of predictive maintenance in roads. Following are key takeaways from the discussion:
Advanced analytics and AI offer an immense opportunity for asset maintenance. The use of digital solutions and analytics has potential to enhance all phases of infrastructure asset maintenance. This topic is at the forefront of industry leaders’ minds: in a poll asking roundtable participants to identify the area where AI’s potential is greatest, the top choices were operations and maintenance (38 percent) and customer engagement (38 percent), followed by tendering and engineering (13 percent each).
Significant value is at stake. In one real-world example, advanced analytics models were able to detect 25 percent of infrastructure failures for a large European rail operator by inspecting just 10 percent of assets.
Further impact and additional use cases lie on the horizon. The digital asset management tools that now enable predictive maintenance could also be scaled up to enable safety and connectivity (for example, connected or autonomous vehicles) in the future. And solutions can be modified across a broader portfolio: road assets may differ from airports, said one speaker, but the foundations of the solution are similar, “so you can start from a single solution and adapt it.”
Data access and quality are a significant challenge. Despite these potential benefits, participants agreed that these technologies present considerable challenges, particularly in terms of data access and quality. One participant remarked, “In our companies, we have many different [types of] software, with each of them having their own database. Sometimes we cannot access the database. [This can come as a surprise], given we are accessing [what we believe] belongs to us. But in fact, it doesn’t.” Another attendee added to this: “The data quality and data structure are very important in this moment. We are building our data lake, and [managing the structuring of data] is consuming, and it affects all the [procedures we are undertaking] in developing new asset management platforms.”
Aside from substantive discussion of data management as a challenge, participants cited other hurdles, including data privacy, implementation failure, defining ROI, and business continuity. “We are constrained, considering the current level of investment, that overall we can’t overcome,” said one participant about their organization’s existing initiatives in predictive maintenance. “But we are confident on the benefit and have a detailed business case to come.”
Change management is essential for the “last mile” of effective implementation. One leader commented, “Data availability and structure are important, but also change management to create trust in using these tools.” Several participants mentioned the cultural shift required for organizations to become data driven as one of the biggest challenges in leveraging new technologies. Making decisions based on data—even if it may run counter to previous ways of working or technical know-how—creates new complexities, including cultural resistance or reliability concerns, for today’s road workforce.
Involving all stakeholders in the design and implementation of AI is key from the beginning to avoid misalignment. One participant described his organization’s efforts to generate awareness of AI across the workforce; these included quarterly updates and initiative ambassadors who could share success stories and practical use cases.
Implementation of predictive maintenance technologies could be business line driven. Several participants agreed that the implementation of predictive analytics or AI technologies should be overseen by business lines, in order to ensure the solutions meet actual operational needs. While technical expertise is crucial, the actual application of the technology needs to be guided by those who understand the business challenges at hand.
One participant stated, "For me, it will really be the business line because we don't speak the same language. Our goal for innovation transformation is to build a bridge between operational needs and the enterprise. For [the latter], [deploying digital tools is] basic, it’s easy, and they don’t understand why these guys don’t want to do it.” The deployment of technologies must be driven by the business need, pain point, and clear definition of the desired outcome.
To learn more about GII’s Europe Roads roundtable series or to request information on our next roads convening in November, contact GII@mckinsey.com.