How AI-enabled KPIs can help leaders align their companies—and get better results

This post was originally published on this site

https://content.fortune.com/wp-content/uploads/2023/11/GettyImages-1500238217-e1699403903426.jpg?w=2048

A vital question for organizations is: How do they measure success? As companies grow larger and more complex, determining what metrics to use to evaluate performance gets harder. 

Traditionally, defining key performance indicators, or KPIs, has been the job of senior executives, who relied on their own judgement, intuition, and experience. But legacy KPIs often score performance on suboptimal or even wrong measures. As companies amass ever-larger, more diverse sets of data, legacy metrics based primarily on human judgment will be less and less likely to align performance dynamics with desired outcomes. KPIs need to become smarter.

AI can help by allowing companies to use their own data to better understand what drives performance. In the process, AI can also change how organizations measure, analyze, and align performance, replacing static, legacy metrics with dynamic, smart KPIs that offer more detailed and accurate descriptions of what is actually going on in a business and what is likely to happen next.

To understand how executives are using AI to improve strategic measurement and outcomes, and how their organizations have adapted to AI-enabled KPIs, Boston Consulting Group (BCG) and MIT Sloan Management Review teamed up to conduct a global survey of over 3,000 managers representing more than 25 industries across 100 countries. We also conducted 17 executive interviews to gain greater context and insight into the experience of individual firms using AI to transform their KPIs.

Our research found that leaders who use AI to prioritize, organize, and share KPIs see improved alignment across units or functions, which in turn drives better overall results. Smarter KPIs can operate as an organizational GPS of sorts, streamlining decision-making and impetus across teams. But how do companies use AI to create and manage new smart KPIs? 

Strategic alignment with AI-enabled KPIs

Our executive survey found that the use of AI-enabled KPIs strongly impacts three dimensions of alignment: 1) Teams are more likely to agree on which KPIs to prioritize; 2) KPIs interlinked across an organization are able to be optimized as an ensemble, rather than in isolation; and 3) teams are more likely to share information when needed, boosting both accountability and alignment. Let’s explore each dimension.

Prioritization. Our survey found that companies that reported using AI to prioritize their KPIs were 4.3 times more likely to say they have more alignment between functions than those not using AI. Teams are often burdened with jumbles of different KPIs that say different things, especially as operations become more complex. Prioritizing KPIs is essential to preventing wasted effort and resources. AI-driven models, as demonstrated by industry leading firms, help prioritize KPIs by algorithmically identifying which have the greatest impact on the desired business outcomes.

Maersk, the Danish transportation, shipping, and logistics company, offers a good illustration of the challenge of prioritizing competing performance indicators. The firm sought to determine how best to evaluate performance: speed (loading and unloading ships or trucks as quickly as possible) versus reliability (managing the loading process to keep to a reliable schedule). 

The company’s port managers argued for speed as the best performance measure on the basis that this would increase throughput—even though additional equipment would be needed to handle the increased pace, which increased short-term costs. Using AI, however, Maersk’s data team, came to a radically different, counterintuitive conclusion: Going slower improved system-wide outcomes. Again using AI, they determined an optimal throughput metric for loading and unloading that was slower than the port managers’ experience suggested.

To build confidence in the AI-powered analyses, Maersk developed a model to test each approach’s impact across their value chain. The company discovered that going faster at one port led to bottlenecks elsewhere, negating any overall productivity gains. In contrast, keeping to a reliable (if slower) schedule resulted in more on-time arrivals, reducing costs. The complexity of these interrelated variables proved hard for human judgement alone to decipher, while AI was able to identify and explain the most useful (“slower”) performance metric. Prioritizing reliability over speed, AI facilitated measurably better organizational alignment for measurably better outcomes.

Ensembling. Metrics that companies, teams and individuals work towards are often interlinked and should not be viewed or optimized in isolation. An IT consulting firm looking to respond to proposals, for instance, targets maximizing the speed with which it can staff a new project while ensuring fitness for purpose. These two objectives—speed versus quality—don’t call for prioritization per se, but rather joint optimization. Using AI, the company can more accurately predict its “win probability” by analyzing the hiring company’s past records and current projects underway. This, in turn, guides resource allocation (including staff time) to high win probability projects.

Ensembling is another area where AI’s pattern-recognition capabilities typically outsmart human judgment and intuition. For example, Pernod Ricard, the $10 billion global spirits brand, uses AI to balance two oftentimes competing strategic priorities: increasing profit margin and increasing market share. AI is able to assess, weight, and deliver insight on how commercial and marketing investments that improve profits also influence market share objectives—and vice versa. In the past, each of these KPIs would be siloed: The finance function focused on profitability, while sales and marketing emphasized market share. Pernod Ricard is now able to dynamically balance its pursuit of profitability and market share, both strategically and operationally, thanks to ensembling its AI algorithm to evaluate overall impact.

Sharing. Dedicated teams and functions often end up with dedicated and siloed KPIs, which hurts overall performance. As one executive remarked, “We need to do more to share KPIs. … What are the right KPIs to share to ensure that one thing isn’t counterproductively overriding the other?” Our survey revealed that organizations using AI to create shared KPIs across teams say they are five times more likely to see improved alignment and three times more likely to be agile and responsive than organizations that do not use AI to share KPIs. 

The essential insight is that AI is able to identify metrics across organizations that require shared accountability. Making KPIs clear and easily accessible also promotes sharing and fosters data-driven conversations across teams. Sanofi has done precisely that through its PLAI app, which uses AI to help analyze, process, and present the numbers that make the most sense to specific audiences within the company. By offering visibility into enterprise-wide performance, PLAI creates a single source of truth that helps people see where they are and what needs to be done. 

Governing AI-enabled KPIs

Adopting AI-enabled performance indicators is not as simple as flipping a switch. Companies pursuing algorithmic innovation must take three key steps: 1) Get company data in order; 2) build organizational constructs to oversee and orchestrate KPI/AI co-development; and 3) strengthen their culture of data-driven decision-making.

Data. Executives we surveyed—without exception—emphasized that clean, trusted data is crucial for KPI transformation. Generating clean data, however, has been notoriously tricky for companies. Organizations getting their data in order for AI-enabled KPIs must focus on two factors: 1) ensuring that systems exist to generate the necessary data; and 2) setting up the data architecture to facilitate the production of KPIs.

General Motors chief data and analytics officer Jon Francis told us that data strategy is of paramount importance because it builds trust in the metrics among employees and executives alike. Thinking through the telemetry, measurement plans, and instrumentation is vital to ensure that data is flowing and can supply the KPIs needed. A good data strategy makes the production of AI-enabled KPIs routine. That means data should be co-located, reporting of metrics should be automated using AI, and teams should not require additional technical support to view and manage their KPIs.

Pernod Ricard’s journey to enable AI interventions highlights the centrality of the data challenge. The company determined it needed three years of weekly sales data, but found almost 80% of that necessary data was external. Collating this data took an extensive manual effort. That made it particularly important to create a business case for the effort to build internal consensus among employees on its value and get the buy-in needed. 

Organizational constructs. Organizations that have successfully introduced AI-enabled KPIs have often been anchored by a team or group of teams dedicated to taking a transversal view. The precise approach may vary, but assigning responsibility to one dedicated entity goes a long way in sustaining a successful KPI transformation program.

Schneider Electric, for example, created a dedicated performance management office to maintain oversight of KPIs deliberately positioned in the governance team to ensure a neutral, cross-functional perspective. To help senior leadership focus on the most important information amid numerous KPIs, the management office updates them on what is most relevant in driving business performance. Singaporean bank DBS, on the other hand, took a more disaggregated approach, incorporating cross-functional squads to analyze process drivers, optimize them, and present a common, focused view to all squad members rather than having thousands of metrics.

Culture. Organizations have historically valued experience and intuition over data in making decisions. As one interviewee observed, the usual pushback against data-driven decision making is that company leaders were paid millions for their gut instincts. That’s why the shift to being data-centric and open to AI-led interventions starts with the leaders themselves.

Sanofi is working to reboot how its executives think. “The most senior 150 leaders are being trained in bootcamps to become more data-centric, more information-seeking, to ask the right questions and really be more digitally savvy in the way they formulate their needs,” Sanofi chief digital officer Emmanuel Frenehard told us. “Our goal as part of our cultural change is that these 150 people, when they leave, they’re far more inclined to adopt that data-centric view. We are making sure that the core of our leadership in the company is trained on how to use the next generation of AI-driven KPIs.” This top-down approach, along with the company’s data-driven conversations catalyzed by the PLAI app, is helping drive cultural change at Sanofi.

Conclusion

Driving strategic alignment within their organization is an increasingly important priority for senior executives. AI-enabled KPIs are powerful tools for achieving this. By getting their data right, using appropriate organizational constructs, and driving a cultural shift towards data-driven decision making, organizations can effectively govern the creation and deployment of AI-enabled KPIs. Doing this will help them prioritize, ensemble, and share KPIs more effectively—the first steps in achieving the goal of strategic alignment.

***

Read other Fortune columns by François Candelon

François Candelon is a managing director and senior partner at BCG, and the global director of the BCG Henderson Institute. You may contact him at Candelon.Francois@bcg.com. 

Michael Chu is a partner and associate director, data science, at BCG X.

Gaurav Jha is a consultant at BCG, and an ambassador at the BCG Henderson Institute.

Shervin Khodabandeh is a managing director and senior partner at BCG.

David Kiron is the editorial director, research, of MIT Sloan Management Review and program lead for its Big Ideas research initiatives.

Michael Schrage is a research fellow with the MIT Sloan School of Management’s Initiative on the Digital Economy. 

Some of the companies featured in this column are past or current clients of BCG.