Service value chains are prime candidates for accelerated digitalization and the integration of artificial intelligence (AI) or generative AI (gen AI) technologies. Many service-oriented companies possess vast stores of customer-relevant data, such as asset history, Internet of Things (IoT) data, technical publications, maintenance manuals, and service requests, which present a rich resource for data mining and deriving actionable insights. The inherent variability and complexity of services make them especially suitable for the application of AI techniques, which can predict outcomes with a precision that surpasses human capabilities.
Dreaded downtime
Ascendum, a Portuguese multinational machinery provider, distributes 25 brands of machinery for the construction, infrastructure, haulage, and agricultural sectors. At the core of Ascendum’s operations, a team of field agents helps maintain and repair machinery, ensuring productivity and minimizing downtime.
Downtime is a major drag on business performance. Indeed, McKinsey analysis of the performance of more than 50 B2B organizations over a 15-year period found that those with a high service focus generated 1.7 times the total shareholder returns (TSR) of those that focused mainly on products. At Ascendum, IT and business development director Rui Galamba says, “Each hour of downtime for our customers costs in the range of $5,000 to $12,000 because it’s very asset-intensive work.”
But keeping Ascendum’s fleet of highly specialized machines running smoothly is hard—and getting harder. Machinery is becoming more complex and specialized, making it difficult to identify a customer’s specific model and issue at hand. Agents would routinely search through a slew of complex, unstructured information in multiple formats (including many thousands of PDFs) and databases, which are frequently updated, making the right information even harder to find. Indeed, simply diagnosing an issue becomes a logistical challenge, given the sheer volume of data and information a field agent must comb through to determine what steps need to be taken.
Operational improvements represented a significant opportunity for Ascendum to improve customer experience and add value, and therefore an opportunity to improve its own operational model.
The secrets of success
Why did this transformation and change program succeed, when so many others achieve only a fraction of the anticipated value?
First, the scope of the initial work was clearly delineated. The planners identified some 30 areas where gen AI could be applied to the repairs process. One use case involved using gen AI to help agents immediately pinpoint equipment repair instructions from a large body of technical documents spread across systems. This enabled field technicians to repair machines as quickly as possible.
A second success factor was the team. Ascendum led the transformation but partnered with McKinsey and Salesforce. The expertly guided process of strategic collaboration and technological integration was a winning combination, delivering results and value for the company and its customers.
Finally, Ascendum maintained a relentless focus on customer service. The team framed their work around the objectives of enhancing first-time issue resolutions by pinpointing repair instructions for field technicians.
The data imperative
More widely, teams and businesses approaching transformation projects in the operations and field service domains need to prioritize the quality and availability of the right data. Introducing gen AI to existing systems has significant implications for the technology stack, which needs to be integrated, scalable, and secure. This may involve upgrading existing systems to ensure that legacy systems can support gen AI capabilities and building new infrastructure that can meet the data requirements of advanced gen AI use cases.
These requirements involve a broad set of data sources and data types, including the following:
- process data: documentation for maintenance, troubleshooting, key failure modes and effects (FMEA) analyses, along with standard operating procedures and work instructions
- IoT data: sufficient data from sensors on equipment to monitor key parameters and predict abnormal conditions
- equipment history: data on equipment configuration and configuration changes, past failures, and maintenance actions
- supply chain data: parts availability, inventory locations, shipping times, and vendor lead times
- personnel data: skills, availability, and location of field service teams and support staff
Organizations need to ensure that this data is available, accurate, and accessible. At the same time, they must protect sensitive information and intellectual property, utilizing AI models to prevent data theft or illegitimate use. That calls for effective data governance with strictly enforced standards for data collection, clear labeling, and robust cybersecurity measures.
The work and results
Quantum Black built an AI-powered search engine for Ascendum within Ascendum’s Salesforce platform, with a user interface build and user acceptance program running in parallel. The results are wholly positive: Besides achieving the primary aim of speedier diagnosis through gen-AI-assisted streamlining of technical information for field agents, the initiative has improved Ascendum’s customer relationships and growth prospects. On the ground, customers experience fewer disruptions, and for construction, industrial, and infrastructure projects, the improvement is a tremendous source of value.
Project legacy and learnings
The Ascendum project demonstrates that innovation through gen AI can deliver transformational top- and bottom-line improvements. It also shows the benefits of putting together and investing in a team that has the right expertise in key areas and can achieve impact at scale and on a favorable timeline.
The project laid the groundwork for ongoing innovation in field service excellence through gen AI. It illustrates how introducing the right expertise and partners to operations can transform a business from the inside out.
Getting started with gen AI in services
Generative AI holds significant potential for aftermarket and field services. Before getting started, companies should consider three key questions:
- What are your service goals? Define what you want to achieve and decide what is critical.
- Where can generative AI have the most impact? A series of interconnected AI use cases is more likely to drive substantial change and help meet your strategic objectives than isolated experiments.
- Does your organization have the necessary digital foundations? These include robust infrastructure, reliable data, and effective governance.
Addressing these questions up front will help your organization build a solid foundation for next-generation B2B services and capture value that benefits customers, employees, and shareholders.