The Art of AI Maturity | Accenture

The art of AI maturity Advancing from practice to performance From insights to action, the path to extraordinary value starts here.

Contents AI maturity: AI maturity: AI Achievers How AI Practice Appendix Why it matters What it is advance from Achievers makes practice to master their progress performance craft 05 08 13 17 29 35 The art of AI maturity—Advancing from practice to performance 2

Executive summary In fewer than 70 years, artificial intelligence (AI) has evolved from a scientific concept to a societal constant. Computer scientist John McCarthy Despite these ever-expanding use cases, This journey to AI maturity has been in coined the term “artificial intelligence” when it comes to making the most of AI’s high gear for years. Pre-pandemic (2019), in 1955, proposing that, “every aspect of full potential and their own investments, AI Achievers already enjoyed 50% greater learning … can in principle be so precisely most organizations are barely scratching revenue growth on average, compared 12% described that a machine can be made to the surface. with their peers. And in 2021, among simulate it."1 executives of the world’s 2,000 largest of firms have In fact, only 12% of firms have companies (by market capitalization), advanced their Today, so much of what we take for advanced their AI maturity enough to those who discussed AI on their earnings AI maturity enough granted in our daily lives stems from achieve superior growth and business calls were 40% more likely to see their machine learning. Every time you use a transformation, according to Accenture’s firms’ share prices increase—up from 23% t o achieve superior wayfinding app to get from point A to extensive analysis of approximately 1,200 in 2018, according to analysis by performance and point B, use dictation to convert speech companies globally. We call them the Accenture. growth. to text, or unlock your phone using face “AI Achievers.” ID ... you're relying on AI. And companies across industries are also relying on—and Another 25% of firms are somewhat investing in—AI to drive logistics, improve advanced in their level of AI maturity, while customer service, increase efficiency, the remaining 63% (the majority) are still 63% empower employees and so much more. mostly testing the waters. of firms are still testing the AI waters. The art of AI maturity—Advancing from practice to performance 3

What do AI Achievers do differently? While there’s clearly a science to AI, our findings demonstrate there is also an art to AI maturity. Achievers are not defined by the sophistication of any one capability, but by their ability to combine strengths across strategy, processes and people. Here are five ways AI Achievers master their craft: 1. Their top leaders champion AI as a strategic priority for the entire organization. 2. They invest heavily in talent to get more from their AI investments. 3. They industrialize AI tools and teams to create a strong AI core. 4. They design AI responsibly, from the start. 5. They prioritize long- and short-term AI investments. Our machine learning models suggest that the share of AI Achievers will increase rapidly and significantly, more than doubling from the current 12% to 27% by 2024. In short, advancing AI maturity is no longer a choice. It’s an opportunity facing every industry, every organization and every leader. The art of AI maturity—Advancing from practice to performance 4

The art of AI maturity AI maturity: Why it matters

AI maturity: Why it matters Figure 1: We project that AI transformation will take less time than digital transformation There is a growing consensus that AI is absolutely essential to competitive advantage. So, it’s no surprise that in 2021, 46% The companies leading the way are of CEOs (of the world’s 2,000 largest already seeing the results—42% said that companies by market capitalization) the return on their AI initiatives exceeded 2 mentioned AI on their earnings calls. their expectations, while only 1% said the return didn’t meet expectations. Our survey of over 1,600 C-suite executives and data-science leaders AI, accelerated from the world’s largest organizations found that nearly 75% of companies have integrated AI into their business strategies With early successes building confidence and reworked their cloud plans to achieve in AI as a value-driver, we estimate that AI AI success. transformation will happen much faster than digital transformation—on average, And companies are putting those plans 16 months faster (Figure 1). Source: Accenture Research into practice: Nearly a third (30%) of Note: Our estimate is derived from a natural language processing analysis of investor calls of all AI pilot initiatives are subsequently the world’s 2,000 largest companies (by market cap), from 2010 to 2021, that referenced “AI” scaled to deliver wide-ranging outcomes, and “Digital” in tandem with “business transformation,” respectively. Data was sourced from S&P from accelerating R&D timelines for earnings transcripts. new products to enhancing customer experiences. The art of AI maturity—Advancing from practice to performance 6

The incentive to move quickly is strong. Figure 2: Evolution of companies' AI-influenced revenue share from 2018 to 2024* We found, for example, that the share of companies’ revenue that is “AI-influenced” more than doubled between 2018 and 2021 and is expected to roughly triple between 2018 and 2024 (Figure 2). Given the evidence, it’s easy to see why companies plan to increase and accelerate their AI investments. In 2021, 19% of companies dedicated >30% of their tech budgets to AI development. By 2024, 49% of companies intend to. Note: Color indicates the achieved AI-influenced revenue threshold within each time period. Source: Accenture Research Note: *2024 = projected *Definition of “AI-influenced” revenues: a. Sales of existing products and services made possible through better AI-driven insights on customers, supply chain, channels ; b. Sales of new products and services made possible by human + AI , c. Higher prices through dynamic pricing ML algorithms. These sales include some cannibalization and net new revenues. In contrast, this definition is excluding higher efficiencies in production operations thanks to AI. The art of AI maturity—Advancing from practice to performance 7

The art of AI maturity AI maturity: What it is

AI maturity: What it is If most organizations are racing to embrace AI, why are some seeing more value than others? To uncover strategies for AI success, not only in data and AI, but also in AI maturity measures the Accenture designed a holistic AI-maturity organizational strategy, talent and framework. Fittingly, our analysis itself was culture—to give companies a strong degree to which organizations conducted using AI. competitive advantage. (See pages 36 have mastered AI-related and 37 for key capability descriptions.) We applied machine learning models capabilities in the right to unravel massive survey datasets and This includes foundational AI uncover drivers of AI maturity that would capabilities—like cloud platforms and combination to achieve high have been impossible to detect using tools, data platforms, architecture and more traditional analytical methods governance—that are required to keep performance for customers, (more on the methodology in pace with competitors. It also includes the Appendix). “differentiation” AI capabilities, like shareholders and employees. AI strategy and C-suite sponsorship, Our research found that AI maturity combined with a culture of innovation comes down to mastering a set of key that can set companies apart. capabilities in the right combinations— The art of AI maturity—Advancing from practice to performance 9

The companies that scored best Figure 3: Only 12% of organizations are AI Achievers in both categories are the AI Achievers. AI Builders show strong foundational capabilities and average differentiation capabilities, while AI Innovators show strong differentiation capabilities and average foundational capabilities. Achievers, Builders and Innovators collectively represent just 37% of surveyed organizations—Achievers accounted for 12%, Builders for 12% and Innovators for 13% (Figure 3). A fourth group we’re calling AI Experimenters—those with average capabilities in both categories—make up the majority (63%) of those surveyed. Source: Accenture Research analysis based on a sample of 1,200 companies The art of AI maturity—Advancing from practice to performance 10

For the world’s 2,000 largest firms by AI-enabled remote systems. And the life Figure 4: Levels of AI maturity by industry, 2021 and 2024* market cap, the percentage of Achievers sciences industry will expand its use of was even smaller: 10%. These numbers AI in efficient drug development. Still, suggest that large firms may struggle to there is enormous room for growth in make the large foundational and cultural AI adoption across all industries and shifts needed to become AI Achievers. an enormous opportunity for those organizations that choose to seize it. Taken together, Achievers, Builders and Innovators tend to have more resources For industry laggards like financial (such as technology, talent and patents) services and healthcare, a range of to deliver on their AI visions and to factors may be contributing to their transform their organizations. Examples relatively low AI maturity—including can be found across a wide range of legal and regulatory challenges, industries: healthcare, financial services, inadequate AI infrastructure and a life sciences, utilities, retail, energy shortage of AI-trained workers. and more. AI, applied While industries like tech are currently far ahead in their respective AI maturity, the gap will likely narrow considerably by 2024 (Figure 4). Automotive is betting on a big surge in sales of AI-powered self- driving vehicles. Aerospace and defense firms anticipate continued demand for Source: Accenture Research Note: *2024 = estimated scores. Industries’ AI maturity scores represent the arithmetic average of their respective Foundational and Differentiation index. The art of AI maturity—Advancing from practice to performance 11

AI, applied across industries • One food delivery service uses deep • A large Australian telco deployed AI • In the public sector, Metro de learning to guide drivers to the best to quantify the effectiveness of its Madrid, one of the world’s oldest delivery routes. AI models analyze individual marketing initiatives. The urban rail systems, deployed AI more than 2,000 variables, from firm was able to measure some 4,000 algorithms to sift through mountains the latest food ordering trends to different marketing metrics and, in of data—on everything from air traffic conditions, to make real-time the process, has created a world- temperature at individual stations, recommendations. class marketing performance insights to train frequency and passenger capability, with a range of strategic patterns, to electricity prices—to • A large chemicals and energy firm and tactical applications. The telco is reduce its annual energy intake is using drones and AI-powered using insights gained from Marketing by 25%. computer vision to monitor its Mix Modeling (MMM) to optimize equipment and remote locations. the allocation of marketing spend, • A major US-based beverage bottler The upshot: More frequent messaging and media. used AI to consolidate data sources inspections at lower cost to the and measure the effect of promotions company and fewer safety risks for • A leading solar-panel installer is on different retailers and markets, its maintenance workers. using satellite photos and deep- boosting the bottler’s annual sales learning algorithms to create fully by 3%. • A Middle East-based telco uses automated rooftop-installation plans AI-driven virtual assistants—which and price estimates. In addition to can communicate in different Arab offering end customers an industry- dialects as well as in English—to first ability to self-design their deftly handle some 1.65 million systems, the company expects its AI- customer calls each month. led design efforts to ultimately lower the firm’s sales costs by 25%. The art of AI maturity—Advancing from practice to performance 12

The art of AI maturity AI Achievers advance from practice to performance

AI Achievers advance from Figure 5: AI Achievers outperform in nearly all capabilities practice to performance AI Achievers thrive when it comes to traditional performance metrics. Pre-pandemic (2019), they already enjoyed any one individual capability, but 50% greater revenue growth on average, by their ability to combine strengths versus their peers. And today, they’re 3.5 across strategy, processes and times more likely than Experimenters to people (Figure 5). see their AI-influenced revenue surpass 30% of their total revenues. By comparison, Innovators typically excel at securing senior sponsorship These companies are going above and embrace training for all employees, and beyond, deploying AI solutions to but they lack the foundational capabilities solve problems, spot opportunities and required to support AI at scale. outperform their peers. What sets the AI Achievers apart? Builders, on the other hand, excel at creating data and AI platforms, but Mastery of multitasking they tend to be weaker at cultivating AI fluency and the innovation culture that When compared with all other groups, AI is needed to drive adoption. Achievers demonstrate high performance Source: Accenture Research across a combination of capabilities. They Note: Each square represents one of the 17 key capabilities. The square is filled in when the AI profile is are not defined by the sophistication of outperforming against peers (higher than the average across all companies in terms of % of companies reaching the mature level). The art of AI maturity—Advancing from practice to performance 14

Turning pilots into production Figure 6: Achievers excel at turning AI pilots into production Achievers have largely moved beyond the AI investment “tipping point,” going from experimenting with new AI in isolation to applying AI at scale to solve critical business problems (Figure 6). Achievers are 25% more likely to scale AI pilots across the enterprise compared with Experimenters. Take product development as an example: Procter & Gamble (P&G) uses “explainable AI” algorithms to harness its proprietary data and formulation models and recommend product improvements. If, for example, the company wants to increase the foam in its dishwashing liquid without changing the price, in-house software developers can now ask the explainable AI to recommend a replacement ingredient. If, however, that new ingredient modifies the liquid’s color, developers can then instruct the AI to search for a new ingredient—and so on and so forth. P&G also relies on AI to generate product formulations that are more likely to perform as expected, resulting in less physical testing of new products. The payoff is lower product-development costs, as well as the ability to better Source: Accenture Research tailor products for specific markets and launch them faster. Note: Score 0-100, ranging from 0 = AI use case not started, 50 = AI use in early stage, 100 = having AI programs in place for full productization. The chart shows the average scores for AI use cases of different functions, between Achievers and other firms. Those differences are statistically significant after controlling for industry, geography and company size; see Appendix for more details. The art of AI maturity—Advancing from practice to performance 15

Focusing beyond firms that managed to reduce emissions financial metrics from their operations, 70% used AI to achieve those reductions. Likewise, of the Achievers also develop strong surveyed companies that made strides relationships with customers—by in measuring and disclosing their carbon building trust, reducing churn and footprints more transparently, 75% used AI boosting the quality and safety of to make such progress. offerings. Our stakeholder performance model showed with high statistical One US-based utility company conducts significance that Achievers score 8% remote monitoring of its extensive grid higher than Experimenters on customer infrastructure via satellites, drones and experience (see Appendix for more other surveillance tools. With the help of on methodology). advanced analytics, machine learning and computer vision, the company is able to Additionally, they double down on their quickly identify and prioritize areas for commitment to sustainability by, for maintenance, improve public safety and instance, rigorously measuring and mitigate the effects of climate change. reducing their greenhouse gas emissions, consuming water and other natural The stakeholder performance resources more economically and model revealed the value-creation gap using AI responsibly. between Achievers and other companies is significant when it comes Accenture’s Sustainable Technology to sustainability. survey of more than 500 multinational companies found that of the surveyed The art of AI maturity—Advancing from practice to performance 16

The art of AI maturity How AI Achievers master their craft Five success factors

How AI Achievers master their craft It’s worth noting that the potential for AI-mature organizations will evolve along with the technology itself. High performance today will ultimately become business-as-usual tomorrow. Today’s AI Achievers have set the standard and are poised to remain leaders. While there is clearly a science to AI, they’ve shown us there is also an art to AI maturity. They have demonstrated that excellence in areas like vision and culture are just as critical as algorithmic integrity. Our research uncovered five key success factors for AI Achievers. The art of AI maturity—Advancing from practice to performance 18

Success Factor 01 Champion AI as a strategic priority for the entire organization, with full "In the last five years, we sponsorship from leadership started to use AI as one of our Companies can create strong AI And for the CEOs of Achievers, main drivers in the business. strategies, but unless those strategies creating a culture of innovation is itself Five years ago, AI was not receive enthusiastic support from the CEO a deliberate, strategic move—one that is and the rest of the C-suite, they’re likely to used as a vehicle for experimentation critical at all … [but today] it’s flounder, competing with other initiatives and learning across the organization. becoming critical. I would say, for attention and resources. In fact, 48% of Achievers embed between 0 and 10, [AI] has Achievers are more likely to have formal innovation in their organizational senior sponsorship for their AI strategies: strategies, while just 33% become something like an 8."3 we found that 83% of Achievers have such of Experimenters do. sponsorship, while only 67% of Builders and just 56% of Experimenters have it. CEO of a German automotive parts and equipment manufacturer Our research also suggests that the best AI strategies tend to be bold, even when they have modest beginnings. Bold AI strategies, in turn, help spur innovation. The art of AI maturity—Advancing from practice to performance 19

For instance, Lendlease Digital (part of We found that 16% of Achievers are multinational Lendlease Group) hopes already using platforms that allow workers to produce architectural blueprints for to easily pose questions and share ideas buildings using generative design and AI, with colleagues across the company— then use those blueprints to manufacture compared to 4% of Experimenters. That actual buildings in factories—fitting number will only grow as these companies together all the pieces like LEGO sets. expand their pools of AI talent. The company’s bold vision starts at the top, led by the CEO of Lendlease Digital, William (Bill) Ruh. To encourage such end-to-end innovation, Achievers implement systems and structures that help employees showcase 83% their innovation experiments and seek constructive feedback from leadership. of Achievers have For instance, Achievers tend to be the first CEO and senior to embrace new tools that encourage their employees to experiment and innovate. sponsorship. The art of AI maturity—Advancing from practice to performance 20

Success Factor 02 Invest heavily in talent to get more from AI investments With a clear AI strategy and strong CEO This makes it much easier to sponsorship, organizations are more likely scale human and AI collaboration to invest heavily in creating data and AI and ensure that AI permeates fluency across their workforces. While AI the organization. proficiency must start at the top, it can’t end there. Nearly half (44%) of Achievers have employees with consistently We found, for example, that 78% of high AI skills competencies, while Achievers—compared with just 56% of Innovators (33%) and Experimenters (30%) Builders and 51% of Experimenters— have significantly fewer such employees, have mandatory AI trainings for most on average. Furthermore, Achievers have employees, from product development employees with higher competencies in engineers to C-suite executives. almost all data- and AI-related skills. Because Achievers prioritize efforts to build AI literacy in their workforces, it’s no surprise that their employees are also more proficient in AI-related skills. The art of AI maturity—Advancing from practice to performance 21

Achievers also develop proactive AI Japanese e-commerce giant Rakuten talent strategies to stay at the forefront of established an “AI Promotion Department” industry trends. In addition to hiring, this in 2016 to accelerate efforts to infuse could mean partnering with or acquiring AI into the company’s 70+ diverse specialist companies to fill critical roles businesses. By 2018, the department (such as data or behavioral scientists, helped turn more than 30 AI pilot 5 social scientists and ethicists). It also projects into successful offerings. means having a plan to get these diverse, 78% multidisciplinary workers to collaborate, And a leading Southeast Asian oil and create and sustain maximum value from gas firm built an AI-powered, “gamified” of Achievers have the company’s data-science capabilities. learning platform to expand employees’ digital fluency. It also created a cloud- mandatory AI trainings In 2018, US utility Exelon established an based performance reviewer that for most employees, from Analytics Academy. This training upskilled scrutinized a decade’s worth of employee employees like Jeffrey Swiatek, who, at the data to recommend workers best suited product development age of 41, transitioned from his longtime for various digital roles. The innovation engineers to C-suite role as a maintenance worker into a saved the firm’s HR department significant executives. higher-paying position as a quantitative time filling positions. It also reduced engineer. Swiatek has since used his scope for managerial bias in promotional training to write predictive software that decisions and helped workers assess and saved Exelon an estimated $1 million over close digital-skills gaps. 4 eight years on equipment maintenance. The art of AI maturity—Advancing from practice to performance 22

Success Factor 03 Industrialize AI tools and teams to create an AI core Another priority for Achievers involves learning engineers, data scientists, data- building an AI core: an operational domain experts and systems engineers. data and AI platform that taps into companies’ talent, technology and data To build AI cores, Achievers harness ecosystems, allowing firms to balance the power of internal and external data, experimentation and execution. An AI making that data trustworthy and storing core helps organizations productize their it in a single enterprise-grade cloud AI applications and integrate AI into other platform—complete with appropriate applications, which makes differentiation usage, monitoring and security policies. with AI more seamless. To extract value from their data quickly An AI core also works across the cloud and effectively, Achievers are also continuum (e.g. from migration to 32% more likely, on average, than innovation), provides end-to-end data Experimenters to either develop custom- capabilities (foundation, management built machine learning applications or and governance), manages the machine work with a partner that offers solutions- learning lifecycle (workflow, model as-a-service. Achievers are also more likely training, model deployment) and than Innovators to use AI for innovation, provides self-service capabilities. AI tapping into readily available developer cores are, in turn, managed by dedicated networks that can swiftly productionize interdisciplinary teams of machine and scale successful pilots. The art of AI maturity—Advancing from practice to performance 23

A major insurer used to rely on its Thanks to the organization’s increased employees to manually administer investment in AI-savvy talent, its business claims—a tedious process that cost department now receives new, AI- over $500 million annually—even powered apps within five months of before the firm’s billion-dollar payouts to initiating their development—compared policyholders. The company explored with an 18-month wait, on average, before native cloud-storage systems and AI the digital factory was built. More broadly, more aggressively, with the goal to store, by 2025, the company expects its digital analyze and track images and other factory to boost its bottom line by $1.5 unstructured data to support claims billion annually. processing. By industrializing its AI tools and teams, the company targeted To strengthen their AI cores, Achievers efficiency gains of 5% in its first year alone, often collaborate with external experts with longer-term cost savings of more to stay on top of of breakthroughs in than $100 million annually. science and engineering. In 2020, for example, American Express partnered A European energy company created with the Indian Institute of Technology a digital factory to help empower Madras to create a Data Analytics, Risk and employees to use analytics and AI-driven Technology laboratory at the prestigious 6 insights in their daily jobs. Among other university. Such innovation ecosystems initiatives, the digital factory trains field help Achievers develop AI apps tailored engineers to work with, and improve, specifically to their needs. machine learning models. The factory also provides mandatory data and AI training to all managers, as well as reskilling and upskilling support to the firm’s entire workforce. The art of AI maturity—Advancing from practice to performance 24

Success Factor 04 Design AI responsibly, from the start As companies deploy AI for a growing The ability to demonstrate high-quality, range of tasks, adhering to laws, trustworthy AI systems that are regulations and ethical norms is critical to “regulation ready” will give first movers building a sound data and AI foundation. a significant advantage in the short- and The potential for regulatory changes in long-term, enabling them to attract new many countries makes the challenge even customers, retain existing ones and build more daunting. investor confidence. In a separate Accenture study of 850 Achievers are consciously applying C-suite executives, we sought to gauge responsible AI with greater urgency than attitudes towards AI regulation and assess their peers. Achievers are 53% more likely, organizations’ readiness to comply. on average, than Builders and Innovators to be responsible by design: designing, Nearly all (97%) respondents believed developing and deploying AI with good that regulation will impact them to some intention to empower employees and extent, and 77% indicated that compliance businesses, and to fairly impact is a company-wide priority. Interestingly, customers and society—allowing many organizations view AI regulation as a companies to engender trust and boon rather than a barrier to success. scale AI with confidence. The art of AI maturity—Advancing from practice to performance 25

For companies, the upshot of The initiative has, among other being responsible by design is achievements, produced a practical an improved ability to meet future methodology and toolkit—the first of requirements, better mitigate risks its kind—that offers detailed guidance and create sustainable value for on how to use AI leveraging the FEAT themselves and their stakeholders. principles, i.e. fair, ethical, accountable Even though only 6% and transparent. For instance, the Monetary of the companies surveyed Authority of Singapore (MAS), the Novartis is creating effective country’s central bank and financial accountability mechanisms and risk- had already implemented regulator, recognized the benefits that AI management controls to keep AI-powered can provide to financial firms. Yet MAS was operations and services aligned with the responsible AI practices, 42% also wary of the threat posed to firms and company’s core values. In addition, notes markets by the illegal and/or unethical use Klaus Moosmayer, Novartis’ Chief Ethics, of surveyed companies aspire of AI. To guard against this, MAS helped Risk and Compliance Officer, it’s designed to do so by the end of 2024. launch the Veritas initiative together to, “provide ... a clear framework for with the financial industry, which aims to ethical use of AI—enabling [employees] to support the responsible use of AI in the challenge their own decisions finance industry. 7 and biases.” The art of AI maturity—Advancing from practice to performance 26

Success Factor 05 Prioritize long- and short-term "I’m on a journey that doesn’t AI investments stop with a move to the cloud. To avoid being left behind, most For Achievers, continued investment companies need to aggressively increase largely involves expanding the scope of AI I’m in the middle of rebuilding their spending on data and AI. One reason to deliver maximum impact, while “cross- Achievers get more out of AI is simply pollinating” AI solutions and redeploying an organization shifting from because they invest more in it. resources in the process. being extremely operational We found that in 2018, Achievers devoted As part of its efforts to create a more into an organization that is 14% of their total technology budgets data-driven organization that can offer to AI, while in 2021 they devoted 28%. In customers highly personalized digital transformational." 2024, they plan to devote 34%. service, Walgreens Boots Alliance migrated from legacy databases to Francesco Tinto, global CIO, SVP, Achievers also understand that their AI advanced cloud databases and analytics, Walgreens Boots Alliance investment journey doesn’t have a finish and built some 100 high-value AI products line. There is, they frequently note, no that create detailed customer profiles “peak AI.” These companies know they and help the company better optimize have only scratched the surface of their inventory and prices. The company has AI transformations and that the quality of been committed to its data- and AI-led their investments matters just as much as transformation since 2020 and doesn't the quantity. intend to slow down anytime soon. The art of AI maturity—Advancing from practice to performance 27

The share of AI Achievers will increase rapidly and significantly, more than doubling from the current 12% to 27% by 2024. The art of AI maturity—Advancing from practice to performance 28

The art of AI maturity Practice makes progress

Practice makes progress The concept of using AI to solve business problems isn’t new. By 2019, there was evidence that While Achievers are advanced relative Those who transform will be the scaling AI beyond proofs of concept had to their peers, they’ll set new standards ones whose teams master the art 8 a significant impact on ROI. Then the for high performance as their own of AI maturity, using cloud as the pandemic hit. For many organizations, maturity evolves. enabler, data as the driver and AI enterprise-wide transformation was an as the differentiator. urgent means of survival. For others, it Other organizations should be asking quickly became a catalyst to thrive. questions to assess their own AI maturity. How can AI help you differentiate? To help get started, Figure 7 has some AI Achievers are thriving. Across industries, sample questions for C-suite leaders, "AI, for us, is really the they’ve moved past cloud migration to according to Accenture’s AI maturity innovation. They’ve capitalized on cloud’s assessment. There are also tools available technology that we’re scale and computing power to tap into to help benchmark AI maturity and new data sources and AI technologies that establish clear paths to progress and pushing...We want to make are widely available. But AI isn’t their secret performance. to superior performance. It’s how they’re sure that AI and machine approaching AI that makes them different. As AI technologies become more learning is embedded in They’ve established that AI maturity prevalent, the future of all businesses is is as much about people as it is about going to look very different—some will everything we do." technology. As much about strategy as it lead the change, and some will be is about implementation. As much about subjected to it. Francesco Tinto, global CIO, responsibility as it is about agility. SVP, Walgreens Boots Alliance The art of AI maturity—Advancing from practice to performance 30

Figure 7: AI maturity assessment: sample questions for C-suite leaders Category Key questions • Does your C-suite have clear accountability for data and AI strategy and execution? • How do you identify potential value, and how are business cases prioritized—considering the potential risks and alignment Strategy and Sponsorship with the overall strategy of the organization? • Are you allocating enough delivery resources to build AI products and services in-house, and are you able to get the most out of your ecosystem partners? • To what extent do you have a cloud platform and technology strategy that supports your AI strategy? Data and AI Core • Do you have an effective, enterprise-wide data platform, as well as strong data management and governance practices, to meet business needs? • Are you using data science and machine learning teams effectively across the lifecycle of AI development? • Is your data- and AI-literacy strategy aligned to your business objectives? • To what extent have you prioritized data and AI fluency for senior leaders, business stakeholders and employees across Talent and Culture your organization? • Do you have a holistic talent model to scale, differentiate, retain and develop AI talent (diverse, dedicated teams of machine learning engineers, data scientists, data-domain experts and data engineers)? • How are you institutionalizing a data and AI culture within your organization? • Do you have an enterprise-wide framework to help you operationalize responsible data and AI from principles to practice? Responsible AI • Are you applying a consistent and industrialized responsible data and AI approach across the complete lifecycle of all your AI models? • Are you methodically tracking the evolution of AI-related laws and regulations across the different jurisdictions in which you operate, while anticipating and preparing for future changes? Source: Accenture Research The art of AI maturity—Advancing from practice to performance 31

Meet the authors Research director Sanjeev Vohra Ajay Vasal Praveen Tanguturi, PhD Global Lead – Applied Growth & Strategy Lead and Thought Leadership Research Intelligence Centre for Data & Insights Principal Director Lead – Applied Intelligence Philippe Roussiere Lan Guan Accenture Research Innovation Lead, Cloud First and AI Global Lead – Data & AI The art of AI maturity—Advancing from practice to performance 32

Contributors Research Leads Yuhui Deeksha Xiong Khare Patnaik Research Manager Research Manager Contributors: Rahul Basole, Mariusz Bidelski, Marcin Bodziak, Tomas Castagnino, Joe Depa, Ray Eitel-Porter , Michelle Ganchinho, Michal Hadrys, Andy Hickl, David Kimble, Carrie Kleiner, Andra Najem, Linda Ringnalda, Paridhi Sharma, Joanna Syczewska, Ezequiel Tacsir, Jonathan Thomas, Jakub Wiatrak, and Yingchuan Zhu Marketing + Communications: Kathy King, Alisyn Abney, Alexa Mouta ThThe art oe art of AI maf AI mattuurirityty——AAddvvanancincing fg frroom pm prraactctiicce te to po peerfrfoormanrmanccee 33

About the research Industries 177 131 106 105 102 99 98 98 97 94 93 91 89 86 85 64 97 Countries 158 Revenue (USD) 116 69 207 65 102 887 98 38 111 76 266 93 65 87 173 77 403 42 34

Appendix Survey Economic modeling and data science R represents the level and evolution stronger explanatory power in the first i From August to September 2021, To assess companies’ AI maturity, as well of a company’s AI-influenced revenues model of “sustaining at >10%” than Accenture surveyed 1,615 C-suite as other measures of performance, we (sustaining at >10%, reaching >30%) AI differentiation capabilities; in the executives at 1,176 of the world’s largest took the following steps: With i = company, t = 2021 and t-1 = second model of “reaching over >30%”, companies—present in 16 industries and 2018, X includes controls for industry, AI differentiation capabilities have it headquartered in 15 countries. 1. Identified key capabilities of AI firm size and company location stronger explanatory power. In other maturity (country). words, AI foundational capabilities are Interviews and case studies We sought to understand the key essential to building the necessary We interviewed 25 CEOs, Chief Data capabilities that contribute to reaching The model is a linear probability Lasso foundation for organizations to enter Officers and Chief Analytics Officers. both an “entry” level of AI maturity (i.e. model, a K-fold cross-validation with the AI race. Meanwhile, AI differentiation We also interviewed Renée Richardson deriving at least 10% of revenues from 10 folds performed. capabilities are key for organizations to Gosline (Senior Lecturer at MIT Sloan AI-influenced initiatives from 2018 to reach the next level of AI maturity. School of Management and Principal 2021) and a higher level of AI maturity 2. Defined "foundational" and Research Scientist at MIT’s Initiative on (i.e. deriving more than 30% of revenues "differentiation" capabilities 3. Built the AI maturity index the Digital Economy) and Christine Foster from 2018 to 2021). To do this, we built In our models, we classified We built two indexes that measure (Chief Commercial Officer at The Alan two machine learning models that Capabilities and ∆Capabilities as AI companies’ AI foundational capabilities it-1 it Turing Institute), as well as numerous AI account for more than 80 capabilities foundational capabilities; Capabilities and AI differentiation capabilities, experts at Accenture. Through research that contribute to the two different Interactions are—as the name respectively, as identified by our two it,t-1 and client work, we also developed levels of AI maturity (see box below). suggests—capabilities with interaction, models. An overall AI maturity index over 40 company case studies on AI with strong senior sponsorship is built as the arithmetic average of transformation. R = β +βX + β Capabilities + and a well-defined AI strategy. We both AI foundational index and AI i 0 1 it 2 it-1 β ∆Capabilities + β Capabilities classified these interaction terms as AI differentiation index, which is indicative 3 it 4 Design thinking Interactions + e differentiation capabilities. of their probability of achieving high it,t-1 it We ran a MURAL session with more than AI-influenced revenue. The median 15 senior data scientists to validate our AI From our models, we discovered maturity index of all companies is maturity model. that AI foundational capabilities have 36/100. The art of AI maturity—Advancing from practice to performance 35

4. Constructed AI profiles based on 5. Measured Achievers’ financial • Customer experience reflects with their supplier networks and foundational and differentiation premium how companies strengthen their inventory levels; our measures include capabilities To assess AI Achievers’ financial sales pipeline by developing supplier diversification, supplier risk, The AI foundational capabilities and AI performance, we used data from strong customer relationships; our and inventory management. differentiation capabilities indexes were S&P Capital IQ to build the following measures include consumer trust, then used to construct a matrix. We regression model: Revenue growth = customer churn, product quality 7. Measured the speed of AI β +βX + β AI Achiever + e i used the top quartile as a threshold on 0 1 i 2 i and safety, and an overall customer- transformation vs. the speed of both axes to cluster all the companies (i = company, AI Achiever as the dummy centric mindset. digital transformation from the survey into four groups: variable, and X including controls • Sustainability reflects how To understand how fast companies i • AI Achievers—the top quartile on for industry, firm size, and company companies strengthen their undergo AI transformation compared both foundational and differentiation location). commitment to environmental to digital transformation, we used median maturity index: 64/100 stewardship; our measures include the frequency of mentions of both • AI Builders—the top quartile 6. Measured Achievers’ stakeholder greenhouse gas emissions, terms on companies’ earnings calls on foundational but not on performance ecological management, resource as a proxy. To do this, we performed differentiation median maturity To assess Achievers’ stakeholder use, water and waste efficiency, and a natural language processing index: 44/100 performance in the areas of customer various environmental solutions. analysis of investor calls of the world’s • AI Innovators—the top quartile experience, sustainability, workforce, • Financial reflects how companies 2,000 largest companies (by market on differentiation but not on and supply chain, we built scores from deliver profitable growth and capitalization), sourced from the S&P foundational median maturity index: 0-100 in these respective areas using operate efficiently. earnings transcripts database. (Note: 50/100 data from FactSet, Arabesque, Oxford • Workforce/employee experience Our analysis included 744 companies • AI Experimenters—all remaining Economics, and S&P Capital IQ, which reflects how companies unlock with a consistent history of earnings companies median maturity index: measure companies' performance their workforces’ full potential; our calls during 2010-21.) Finally, we 29/100 against their industrial peers. The measures include compensation, built predictive S-Curve models that difference between Achievers and employment quality, employee estimated the time, henceforth, that it other companies is highly statistically turnover, occupational health would take for 90% of such companies significant (p < 0.01) for customer and safety, and training and to mention the aforementioned terms experience and sustainability. development. on their earnings calls. The following offers more detail on • Supply chain reflects how each area. companies manage risks associated The art of AI maturity—Advancing from practice to performance 36

Key Capabilities Strategy and Sponsorship as a reaction to a need. They’re first- with a partner who offers solutions- keep data clean and trustworthy, and movers instead of fast followers in as-a-service, vs. purchase “off-the- to support decision making at greater 1. Senior Sponsorship: Organizations terms of applying AI for business value. shelf” AI solutions with little-to-no speed and scale. have an AI strategy that is developed customization. by the Chief Analytics Officer, Chief 4. Readily Available AI and ML tools: 9. Data Management and Governance: Data Officer, Chief Digital Officer Organizations work with an ecosystem 7. Platform and Technology: Organizations scale their data or an equivalent. The CEO and the of technology partners to access Organizations apply the necessary management and governance Board actively sponsor and share machine learning models and tools cloud, data and AI infrastructure, practices to increase data quality, accountability for the strategy and to help innovate new products software, self-serve capabilities and trust and ethics across entities associated AI initiatives. and services. industry best practices, and they adopt —e.g. by implementing master data the latest tools available from platform management and ensuring security, 2. AI Strategy: Organizations not only 5. Readily Available Developer and technology partners. compliance and interoperability. have a core AI strategy aligned to the Networks: Organizations tap into an overall business strategy, but they ecosystem of technology partners 8. Experimentation Data—Change: 10. Data Management and Governance— also dedicate tools and tactics to to access developer networks that Organizations improved their use Change: Organizations improved their execute it and continuously track their support the development of new of experimentation data between data management and governance performance against that strategy. products and services. 2018 and 2021, effectively translating practices between 2018 and 2021, into a higher data and AI maturity. effectively translating into a higher data 3. Proactive vs. Reactive: Organizations Experimentation data is the use of and AI maturity. have the resources (such as Data and AI Core internal and external data to design technology, talent and patents) to new models and generate new proactively define and demonstrate 6. Build vs. Buy: Organizations develop insights. To do that, organizations use how AI can create value vs. apply AI custom-built AI applications or work enterprise-grade cloud platforms to The art of AI maturity—Advancing from practice to performance 37

Talent and culture 13. Innovation Culture Embedded: just ML engineers—such as Organizations ensure innovation is part behavioral scientists, social scientists, 11. Mandatory AI Training: Organizations of the day-to-day work environment. and ethicists. enforce AI-specific training programs to They encourage mindsets, behaviors improve AI fluency, which are tailored and routines that all serve as a vehicle Responsible AI for senior leadership and specific for experimentation, collaboration functions, e.g. salesforce, product and learning from ideation to product 16. Responsible AI: Organizations have engineers, etc. They also create development to market launch. an industrialized, responsible approach deliberate opportunities for employees to data and AI across the complete to learn and apply AI in their roles. 14. Innovation Culture Encouraged: lifecycle of their AI models—an Organizations promote and reward approach that can meet changing 12. Employee Competency in AI- innovative mindsets and behaviors regulatory requirements, mitigate risks, Related Skills: Organizations regularly including entrepreneurship, and support sustainable, trustworthy AI. measure the competency level of collaboration and thoughtful employees to determine where further risk-taking. 17. Responsible AI—Change: training is needed to improve overall Organizations have improved their acumen. They measure and build 15. AI Talent Strategy: Organizations have responsible data and AI practices expertise in critical areas like coding, an AI talent strategy—hiring, acquiring, between 2018 and 2021, effectively data processing and exploration, retention—that evolves to keep pace translating into a higher data and business analytics, domain and with market or business needs. They AI maturity. business acumen, machine learning, also have an AI talent roadmap for visualization and more. hiring diverse AI-related roles, beyond The art of AI maturity—Advancing from practice to performance 38

References 1 https://news.stanford.edu/news/2011/october/john-mccarthy-obit-102511. 5 https://global.rakuten.com/corp/careers/topics/engineering3/ html#:~:text=John%20McCarthy%2C%20a%20professor%20emeritus,He%20 6 was%2084 https://economictimes.indiatimes.com/tech/software/american-express-sets- up-data-analytics-risk-technology-lab-in-iit-madras/articleshow/77925793.cms 2 Accenture Research analysis of the world’s 2,000 largest companies by market 7 capitalization mentioning AI in their earnings calls. Formula is based on CEOs of https://www.novartis.com/about/strategy/data-and-digital/artificial-intelligence/ companies that had earnings call in 2020, and CEO was present at the call, and our-commitment-ethical-and-responsible-use-ai CEO mentioned AI. 46% of these CEOs mentioned AI in their earnings calls, in 2021 up from ~35% in 2017. 8 https://www.accenture.com/us-en/insights/artificial-intelligence/ai-investments 3 Accenture Interview 4 https://www.nytimes.com/2020/07/13/business/coronavirus-retraining-workers. html The art of AI maturity—Advancing from practice to performance 39

corp.rakuten.co.jp

novartis.com

accenture.com

About Accenture About Accenture Research About Applied Intelligence Accenture is a global professional services company Accenture Research creates thought leadership about Applied Intelligence is Accenture’s approach to scaling with leading capabilities in digital, cloud and security. the most pressing business issues organizations face. AI for clients by embedding AI-powered data, analytics Combining unmatched experience and specialized Combining innovative research techniques, such as and automation capabilities into business workflows, skills across more than 40 industries, we offer Strategy data science led analysis, with a deep understanding of accelerating time to value with a powerful global alliance, and Consulting, Interactive, Technology and Operations industry and technology, our team of 300 researchers innovation and delivery network that can deploy and scale services—all powered by the world’s largest network of in 20 countries publish hundreds of reports, articles and AI within any market and industry. Advanced Technology and Intelligent Operations centers. points of view every year. Our thought-provoking research Our 710,000 people deliver on the promise of technology developed with world leading organizations helps our and human ingenuity every day, serving clients in more clients embrace change, create value, and deliver on the than 120 countries. We embrace the power of change to power of technology and human ingenuity. create value and shared success for our clients, people, shareholders, partners and communities. Visit us at www.accenture.com. Disclaimer This content is provided for general information purposes and is not intended to be used in place of consultation with our professional advisors. Copyright © 2022 Accenture. All rights reserved. Accenture and its logo are registered trademarks of Accenture. The art of AI maturity—Advancing from practice to performance