Many companies are dabbling in artificial intelligence, but how many are scaling it effectively across their organizations? Probably fewer than you imagine.

To quantify how companies go about scaling AI, we developed a framework for assessing correlations between firms’ financial performance and their AI and foundational capabilities (both observed and stated). Within this framework, we used machine learning to evaluate performance across 30 metrics for 1,511 of the world’s largest publicly traded companies spanning 18 industries. We also surveyed 1,020 CEOs at many of these firms.

We found just 10 per cent of large companies have made AI central to their success. And these leaders are harnessing their strong AI capabilities to help power superior financial results.

The good news for everyone else is that becoming an AI leader is achievable. Drawing on the aforementioned research, along with other empirical studies done by Accenture and lessons from our years of experience helping clients deploy and scale AI, this article highlights five priorities that will help you join the AI leadership ranks.

Your C-suite should be “hands on”

Relatively few top executives have deep expertise in AI. In fact, only 22 per cent of the companies we analyzed have at least half their C-suite and board composed of people with STEM backgrounds.

No wonder it’s tempting for CEOs to step back when it comes to building and scaling AI across the organization—better to let the technical experts run the show, the thinking goes. Yet at AI leaders, the opposite is true: the C-suite takes a hands-on approach.

Indeed, we found that CEOs at such companies actively participate in shaping their organization’s data and AI strategy, such as by setting clear expectations and metrics for measuring progress. These CEOs also spend time educating themselves about AI—and it shows: They’re twice as likely as their peers to have a strong understanding of AI and how it can create value for their enterprise.

At AI leaders, CEOs and other C-suite leaders are  heavily involved in nurturing an organizational culture that prioritizes data and AI. They’re also eager to make that visible, communicating to external stakeholders how their company is using AI to create value; we found that CEOs of AI leaders do this on earnings calls twice as often as their peers do.

One company embracing the concept of AI-led reinvention is Sanofi. The French health-care firm is on track to become the first pharmaceutical company powered by AI at scale. Spearheading its efforts is CEO Paul Hudson, who serves as the key architect of Sanofi’s AI transformation and has been pivotal in making bold moves, investing in AI clusters, and overseeing change-management initiatives.

Sanofi’s reinvention journey is as comprehensive as it is ambitious. The AI strategy touches every part of the enterprise—from drug discovery and R&D to clinical and business operations and patient outcomes. Each area is using or developing AI-powered tools to generate insights that allow Sanofi’s people to make better everyday decisions. For example, the company’s Plai app, developed in partnership with Aily Labs, provides a 360° view of its operations, aggregates data across business functions, generates insights, and enables Sanofi researchers to reimagine how they design and run clinical trials.

Since launching its AI reinvention in 2019, Sanofi has scaled its AI ecosystem and strengthened its AI capabilities across the value chain. As a result, the company has experienced a 20–30 per cent increase in potential target identification in key therapeutic areas, achieved an 80 per cent prediction rate for low-inventory positions within the biopharma supply chain, and created innovative platforms for analytics and drug development.

Build and strengthen your digital core

Companies with a “digital core” can leverage cloud, data, and AI to create the interoperable, secure IT platforms needed to develop new business capabilities and spur growth. These companies intentionally shift from a traditional tech stack—a collection of technologies working in silos—to the digital core, where the foundational components interact seamlessly to adopt and deploy new technologies, such as AI.

Trying to scale AI across your organization without a strong digital core is like driving a sports car with a decrepit engine—you may look good for the first mile, but you won’t go very far. This explains why AI leaders are 2.5 times more likely than their peers to prioritize their digital core.

Whether you’re trying to scale AI as a private or public-sector organization, strengthening your digital core should always be a top priority. The Saudi Data & AI Authority (SDAIA) exemplified this lesson to a great extent. Established in 2019, SDAIA developed a “National Data Bank” to enable government agencies in Saudia Arabia to share data securely and efficiently. To make this goal a reality, SDAIA prioritized the creation of a strong digital core for the National Data Bank, taking care to fortify each of the foundational “layers” of its core (i.e., an infrastructure and security layer, a data and AI layer, and an applications and platforms layer).

Today, the National Data Bank is connected to more than 200 government systems and was also instrumental in establishing Estishraf, an analytics platform that supports planning for policymakers across the country. In 2023, more than 85 government entities benefited from Estishraf’s services, resulting in an estimated 50 billion SAR (about $13 billion) in value delivered for citizens and government agencies.

Reskill your people and reimagine their work

Given the technical IT challenges of embedding AI across an organization, it’s no surprise that many companies tend to overlook the human part. When Accenture surveyed thousands of workers and C-suite executives, for example, 94 per cent of employees said they were ready to learn new skills to work with generative AI, but only 5 per cent of CXOs said their organizations were offering reskilling opportunities to their entire workforce. That’s an enormous gap, especially as ever more working hours (up to 44 per cent in the United States, by our estimate) are affected by generative AI-led automation and augmentation.

AI leaders, though, understand this need and are making big efforts to reskill employees to both do new work and work in new ways with AI. These companies are also being more aggressive in hiring individuals—nearly two times as likely as other firms—who can work with and alongside AI. AI leaders, moreover, realize that to build trust in and expertise around AI, workers must be active co-creators in the scaling of the technology across their organization.

Accenture, for instance, enlisted teams in its sales function in a two-pronged effort to infuse generative AI into their work. The sales teams were first asked to revamp their processes and workflows. The teams then worked with engineering and data science colleagues to design and implement the new AI tools that would support the revamped processes and workflows.

The (self-reported) early returns from such co-creation are impressive: increased productivity, increased confidence, and increased ability to manage stress at work. These positive effects were also greater for people who had been in their role longer—suggesting that experienced employees are especially ready for tools that help simplify and improve their work.

Keep your AI responsible

“Responsible AI” aims to ensure that your algorithms support outcomes that are not only legal and profitable but also ethical and otherwise consistent with your company’s values. Facial-recognition technology that is, say, used to harass political dissidents is not responsible AI. Nor is a judicial-sentencing algorithm that recommends a harsher sentence because of an individual’s skin color.

To their credit, many executives have responsible AI on their radar: in one survey, 39 per cent of respondents told Accenture that their enterprises had experimented with responsible AI pilot projects. Yet even these companies tend to treat responsible AI as simply another box to check for their compliance department.

To avoid this mistake, organizations should explicitly design and build responsible AI from the outset. By steering their AI work in an ethical direction early on—as AI leaders do twice as frequently as their peers—companies reduce the scope for undesirable behavior, while also making it easier for compliance analysts to spot problems that do emerge.

The Monetary Authority of Singapore (MAS), the central bank and financial regulatory authority of Singapore, wanted to help financial services institutions (FSIs) use AI responsibly, while mitigating unforeseen negative consequences. To that end, the Authority established the Veritas industry consortium, which now has more than 25 members. A core team within Veritas—led by MAS and Accenture—co-created a framework and methodologies that enable FSIs to evaluate their AI and data analytics solutions against the principles of fairness, ethics, accountability, and transparency to embed responsible AI into their operations.

Boost your AI investment

“Slow and steady wins the race” is often smart advice in business (and in life). To become an AI Leader, however, companies need to invest less like tortoises and more like hares; generous investment in AI dramatically expands your ability to discover the many ways that AI can improve your business.

This view, grounded in experience, is confirmed by our latest research: AI leaders devote about 4 percentage points more of their tech budgets to AI than other firms. That gap translates into an even wider disparity in AI use cases: AI leaders have 22 per cent more of them, on average, than do other enterprises. Among other benefits, more use cases improve the odds that you’re working on the “performance frontier,” or cutting edge of AI, for your industry. And it generally means that you’re devoting more resources to enriching the experiences of your customers, too.

Consider the insurance industry. Previous research by Accenture found that keeping customers happy is one of the industry’s biggest challenges—a third of claimants said they were not satisfied with their insurer, primarily because they had to wait too long for their settlements.

These claimants, who represent hundreds of billions of dollars in renewal premiums, are thus up for grabs. Savvy insurers have since responded by supercharging their AI-focused tech investments. In 2023 alone, one such insurer spent over $1.5 billion on tech to strengthen the firm’s data, AI, and analytics capabilities. By investing boldly, the company is creating “omni-channel” environments that leverage chatbots, text messaging, and guided scripting to help agents speed up settlement times and keep customers coming back.

As artificial intelligence continues to reshape how business gets done, the benefits of scaling AI—and the costs of not doing it—will continue to grow. By focusing on the five priorities described in this article, your organization can accelerate its own AI-led reinvention.

Acknowledgement: The authors thank David Kimble, Yuhui Xiong, Deeksha Khare Patnaik, Jakub Wiatrak, Sachin Guddad, Paridhi Sharma, and Ammar Mohammed for their contributions to this article.