AI: Built To Scale

Provocative thinking, transformative insights, tangible outcomes AI: BUILT TO SCALE From experimental to exponential Achieve competitive agility

About the authors Ketan Awalegaonkar Robert Berkey Greg Douglass Athena Reilly Managing Director Managing Director Senior Managing Director Managing Director Accenture Applied Intelligence Accenture Applied Intelligence Accenture Strategy Accenture Strategy Ketan leads Strategy & Consulting across Robert has over 20 years of experience Greg leads Communications, Media and Athena helps C-suite leaders address all Accenture Applied Intelligence industry shaping and delivering enterprise analytics Technology within Accenture Strategy. some of their top challenges, including and functional practices. Ketan partners strategy and transformation programs for His role focuses on helping clients changing business models, managing with Fortune 500 C-suite executives and Fortune 500 clients, including value worldwide achieve high performance data volumes, addressing analytics board members to transform their digital targeting, operating model & talent, through profitable growth, accelerated immaturity and transitioning away from and analytics operating models by applying analytic delivery models, data & innovation, organizational agility, and legacy technology. intelligence through design-thinking, technology architecture and employee operational excellence. AI, data strategy and a cloud-based adoption programs. For 20 years, Athena has guided teams in platform ecosystem. Greg has over 25 years of consulting developing actionable strategies and He has co-authored several white papers on experience across the telecom, media, plans to create more value through He teaches AI at both the Kellogg School of analytics transformation including “The technology and retail industries, having analytics and define the optimal Management & McCormick School of Insight-Powered Enterprise” and “Preparing focused on new digital business launches, technology footprint. A frequent Engineering at Northwestern University. for a Data Science Transformation”, and strategic digital planning, business growth commentator on digital trends in national Ketan in based in Chicago, Illinois. co-created Accenture’s Analytics strategies and cost transformation. Greg is and global media publications, Athena is Diagnostic (patent pending). Robert is based in Dallas, Texas. based in San Francisco, California. based in Portland, Oregon. AI: BUILT TO SCALE 2

THE NUMBERS 84% TELL THE STORY of executives believe they won’t achieve their growth objectives A full 84% of C-suite executives believe they must leverage unless they scale AI Artificial Intelligence (AI) to achieve their growth objectives. Nearly all C-suite executives view AI as an enabler of their strategic priorities. And an overwhelming majority believe 76% achieving a positive return on AI investments requires scaling of executives across the organization. struggle with how to scale AI across Yet 76% acknowledge they struggle when it comes to scaling the business it across the business. What’s more, three out of four C-suite executives believe that if they don’t scale AI in the next five years, they risk going out of business entirely. 75% With the stakes higher than ever, what can we learn from of executives believe companies that successfully scale AI, achieving nearly 3x the they risk going out return on investment and a 30% premium on key financial of business in 5 years valuation metrics? if they don’t scale AI AI: BUILT TO SCALE 3

Nail it, then scale it To answer that question, Accenture conducted a landmark global study involving 1,500 C-suite executives from organizations across 16 industries. The study focused on determining the extent to which AI enables the business strategy, the top characteristics required to scale AI, and the financial results when done successfully. The aim: Help companies progress on their AI journey, from one-off AI experimentation to gaining a robust organization-wide capability that acts as a source of competitive agility and growth. Three distinct groups of companies with increasing levels of capability required to successfully scale AI emerged from the research. Proof of Concept Industrialized 01 02 03 Factory Strategically Scaling for Growth In our experience, most companies (80-85%) are Only 15-20% of companies have made this leap. These Very few (<5%) companies have progressed to this point stuck on this path.They conduct AI experiments and companies have journeyed beyond proof of concept to on their AI journey. These companies have a digital pilots but achieve a low scaling success rate and a achieve a much higher success rate scaling AI—nearly platform mindset and create a culture of AI with data low return on their AI investments. Their efforts tend double. And a much higher return—nearly three times and analytics democratized across the organization. to be siloed within a department or team and are their counterparts. As a C-suite priority, these They have scaled thousands of models with a responsible often IT-led. They lack a connection to a business companies have a clear AI strategy and operating AI framework. They promote product and service outcome or strategic imperative. The time and model linked to the company’s business objectives, innovation and realize benefits from increased visibility investment it takes to scale is underestimated, supported by a larger, multi-dimensional team into customer and employee expectations. Our leaving the full potential of AI untapped. championed by the Chief AI, Data or Analytics Officer. research indicates that industrializing AI will enable However, the scaled AI is generally across point competitive differentiation which is correlated with solutions, e.g., personalization. significantly higher financial results. What do Artificial Intelligence (AI) encompasses Pilot: Rolling out a capability with real data, Scale: Extension of the piloted capability we mean? multiple technologies that enable users and processes in a production across the full applicable scope with all computers to sense, comprehend, act, and environment (using a subset of the relevant relevant data, end users, customers, and learn. AI includes techniques such as scope). The purpose is to test how the processes. Purpose is to maximize the machine learning, natural language capability performs with a limited scope and application’s value to the organization. processing, knowledge representation, to make any needed modifications before computational intelligence, among others. expanding to the full applicable scope. AI: BUILT TO SCALE 4

Paying dividends: Proven premium value The C-suite executives surveyed +35%Enterprise Value/ reported positive ROI on their AI Revenue Ratio investments. We dug deeper. Was there any relationship between successfully scaling AI across the enterprise and key market valuation metrics? What was the “premium” for being a leader? Price/Earnings Using survey data combined with publicly available financial data, +33%Ratio our team of data scientists created a model to identify the premium for companies in our sample that successfully scale AI, controlling for various characteristics of the companies. We discovered a positive correlation between successfully scaling AI and three key measures of financial valuation: Enterprise Value/ Revenue Ratio, Price/Earnings Ratio, Price/Sales Ratio. Price/Sales +28%Ratio Companies that were identified as Strategic Scalers realize a success rate of 70% or more in their AI scaling initiatives and a return on their AI investment of 70% or higher. AI: BUILT TO SCALE 5

The great divide When the 1,500 companies in our research were US $110m gap analyzed collectively, US$306 billion was spent on AI applications in the past three years. The ROI gap amongst them was significant. On average, it spanned US$110 million between companies stuck in Proof of Concept and those who have progressed to 1 becoming Strategic Scalers. 86% Strategic Overall reported spend on Artificial Intelligence Scalers initiatives over the past three years Over $500 million 32% $10 million Between or less $101-$500 million Proof of 9% Concept 26% 14% Between $51-$100 million 14% 38% The difference in return on AI investments between companies in the Proof of Concept stage and Strategic Scalers equates to an average of US$110 million. $11-$50 million AI: BUILT TO SCALE 6

Companies strategically scaling AI have nearly 2x the success rate and 3x the return from AI investments vs. companies pursuing siloed proof of concepts AI: BUILT TO SCALE 7

AI’s evolutionary paths to growth Industrialized 03 for Growth • Digital platform mindset and enterprise culture of AI democratizing real-time insights to drive business decisions 02 Strategically Scaling • Clear enterprise vision, accountability, metrics, and governance breaking • CEO focus with advanced analytics and down silos data team solving big rock problems • ‘What if’ analysis enabling improved Proof of Concept • Multi-disciplinary teams of 200+ acquisition, service and satisfaction 01 Factory specialists championed by Chief AI, • Responsible business practices Data or Analytics Officer enhancing brand perception and trust • Analytics buried deep and not a • Able to tune out data noise and focus • Competitive differentiator and value CEO focus on essentials creator driving higher P/E multiples • Siloed operating model typically IT-led • Intelligent automation and predictive • Unable to extract value from their data reporting • Struggle to scale as unrealistic • Catch up on digital/AI/data asset debt expectations on time required • Experimental mindset achieving scale • Significant under investment, yielding and returns low returns AI: BUILT TO SCALE 8

The research revealed three critical success factors that separate those that Roadblocks to have progressed to Strategically Scaling scaling and those still in Proof of Concept. When executives ranked their top challenges for scaling AI, they placed “lack of budget” at the bottom of the list. A possible explanation: AI is a C-suite Strategic Scalers: priority. So, while it may be challenging to decide which initiatives to fund first, the monetary resources, overall, aren’t a problem. 01 Drive “intentional” AI Among the top challenges? The inability to set up a supportive organizational structure, the absence of foundational data capabilities, and 02 Tune out data noise the lack of employee adoption. As the study shows, it’s exactly these aspects where Strategic Scalers outperform their 03 Treat AI as a team sport counterparts in Proof of Concept. AI: BUILT TO SCALE 9

01 Drive “intentional” AI To all intents Creating value from AI requires leaders to anchor and purposes AI in C-suite objectives. Most global organizations today believe passionately in the value of data and analytics. But one life sciences company had been struggling to move from theory to execution in implementing data and insights capabilities Strategic Scalers pilot and successfully scale more initiatives than their across all its business divisions. And they had a Proof of Concept counterparts—at a rate of 2:1—and set longer timelines. vision to scale a collaborative data-powered They are 65% more likely to report a timeline of one to two years to move For Strategic service delivery model to create an internal from pilot to scale. And even though they achieve more, Strategic Scalers Scalers, 8 of 10 marketplace for FAIR (findable, available, spend less. At first glance it may seem paradoxical. But the data indicate scaling initiatives interoperable, reusable) data. that these leaders are more intentional, with a more realistic expectation in terms of time to scale—and what it takes to do so responsibly. are successful Working with data and digital leadership, the multi-functional team designed and delivered a To successfully scale, companies need structure and governance in holistic data and analytics strategy while place. And the Strategic Scalers have both. Nearly three-quarters of achieving immediate value through targeted them (71%) say they have a clearly-defined strategy and operating use-cases in each of the key areas of the business. model for scaling AI in place, while only half of the companies in Proof In addition to architecting and standing up the of Concept report the same. new scalable data and analytics delivery model, they created a model to make data search easier Strategic Scalers are also far more likely to have defined processes and more intuitive and embedded a new data- and owners with clear accountability and established leadership driven culture within the organization. support with dedicated AI champions. Initiatives not firmly grounded in business strategy and lacking a governance construct to oversee The result? The company’s digital and manage are slower to progress. Turf wars break out over who transformation is speeding ahead, powered by “owns” AI and data. And, regardless of the AI platforms used, or the data analytics insights across its business. know-how recruited, misaligned efforts fall flat. AI: BUILT TO SCALE 10

Do the basics brilliantly When it comes to scaling AI, Strategic Scalers do the basics brilliantly. Compared to companies in Proof of Concept, they have a clearly defined strategy and operating model for AI, defined process and owners for measuring value from AI, clearly defined accountability, appropriate levels of funding, Sizing up and flexible business processes with embedded AI. They also scale through reusable assets on platforms, so successive AI programs are 3-5X faster to market at lower spend. the situation The “smaller” companies in our study generated revenues between US$1 and 5 billion a year. The largest had revenues of more than US$30 billion. When it comes to scaling AI, are there any major differences +60% +50% +50% between these two groups of companies? Do the largest companies face lower scaling Clearly defined Clearly defined Appropriate success rates due to their organizational accountability strategy and operating level of funding complexity? Or, quite the opposite, do they for scaling model for scaling achieve higher returns as they untap greater value potential? When we grouped the surveyed companies by size, we found no significant differences in +26% scaling success rate or return on AI +30% +27% investments. So, size is not a factor. It’s all Flexible business about instilling the right AI capabilities and Defined process and Packaged or custom processes with mindset in the organization. owners for measuring built AI applications value from AI for scaling embedded AI AI: BUILT TO SCALE 11

Benefits across the business Optimization in action Companies from around the world and across industries are using AI to change the fabric of what they do and how they do it. Strategic Scalers are more likely to achieve a range of With a continuously evolving competitor and customer benefits including: landscape, a major North American technology company with global reach needed a dynamic integrated approach for its vast and complex supply chain. A mix of open source machine learning tools was used to analyze a range of factors including demand uncertainty, cost drivers, customer relationships, and rate Organization of technological change. Personalized supply chain strategies for Customer Customer effectiveness/ New product four key segments were created with synergies pinpointed to service satisfaction workforce offerings reduce complexity and maintain economies of scale. productivity The result? A personalized supply chain transformation with AI at the core driving new efficiency and effectiveness: 30% faster Product/ speed to delivery with 2X more accurate forecasting. Operating Customer Customer brand income up 45% with a reduction in manufacturing and freight acquisition trust perception costs of over 25%. Product availability improved over 35%, while overall supply chain was streamlined with over 90% fewer product configurations. Spend Working more accurate management capital efficiency utilization 2X forecasting increase in 45%operating income Flexible business processes AI: BUILT TO SCALE 12 with embedded AI

02 Tune out data noise Strategic Scalers recognize the importance of managing data. Ninety percent of the data in the world was created in just the past 10 years. One-hundred and seventy-five zettabytes of data will be created by 2025. Yet after years of collecting, Strategic Scalers strongly agree: storing, analyzing, and reconfiguring troves of information, most organizations struggle with the sheer volume of data and how to cleanse, manage, maintain, and consume it. Strategic Scalers tune out “the noise” surrounding data. They recognize the importance of My organization business-critical data— identifying financial, marketing, consumer, and master data as recognizes the priority domains. And Strategic Scalers are more adept at structuring and managing data. They have invested heavily in data quality, data management, and data governance importance of our core frameworks on the cloud. And they have clear operating models for generation versus consumption of data. The research shows they are much more likely to wield a larger, data as the foundation to more accurate data set (61% versus 38% of respondents in Proof of Concept). And 67% of Strategic Scalers integrate both internal and external data sets as a standard practice scaling AI. compared to 56% of their Proof of Concept counterparts. What’s more, they use the right AI tools—things like cloud-based data lakes, data vs engineering/data science workbenches with model management and governance, data and 54% 37% analytics marketplaces and search—to manage the data for their applications. From creation to custodianship to consumption. Strategic Scalers understand the importance of using Strategic Companies in the more diverse datasets to support initiatives. Scalers Proof of Concept stage AI: BUILT TO SCALE 13

Masters of their data Tapping into data Strategic Scalers invest in a data foundation that enables them to scale AI. With beer sales declining and competitors drinking up market share on all sides, brewers are under increasing pressure to find new growth. One global brewer found a solution by tapping into new intelligence. The latest machine learning techniques helped overcome data veracity issues and develop more accurate forecasting models, improved consumer and customer segmentations, and enhanced sales incrementality. Their new advanced analytics capabilities are scaled to over 100+ global datasets, including sales and forecast data, social media, trade spend, customer and product master data, even weather data. They get actionable Proof of Concept Strategic Scalers data-driven insights in front of their key business decision makers, from commercial intelligence to sales and marketing, at unprecedented speed and scale. The result: A return of four times the investment in the first year alone. AI: BUILT TO SCALE 14

03 Treat AI The “A” and “I” as a team sport in team A leading global convenience store chain with The effort of scaling calls for embedding $60B in revenue and 16,000 locations sought to gain a more competitive position to stave multi-disciplinary teams throughout the organization off more agile rivals. in addition to having sponsorship from the top. They leveraged data science to price their products more competitively to better match customer demand across global markets. Leveraged machine learning and automation The effort of scaling calls for embedding multi-disciplinary teams to increase pricing frequency. And throughout the organization—teams with clear sponsorship from the top established virtual agents to interact with ensuring alignment with the C-suite vision. For Strategic Scalers, these their global category management teams to teams are most often headed by the Chief AI, Data or Analytics Officer. 92% drive adoption of the new pricing approach. They’re comprised of data scientists; data modelers; machine learning, All of these changes were made possible data and AI engineers; visualization experts; data quality, training and of Strategic thanks to multi-disciplinary teams with skills communications specialists. Scalers leverage in areas like data engineering, visualization, multi-disciplinary data quality, and human-centered design. It’s a lesson Strategic Scalers have learned well. In fact, a full 92% of them teams The initiative is expected to deliver an leverage multi-disciplinary teams. Embedding them across the organization is expected US$300 million in gross profit not only a powerful signal about the strategic intent of the scaling effort, it uplift annually once fully scaled. also enables faster culture and behavior changes. In contrast, those still in Proof of Concept are more likely to rely on a lone champion within the technology organization to drive AI efforts. AI: BUILT TO SCALE 15

All hands on deck As companies journey toward maturity in scaling AI, new skill sets emerge as employee base has an opportunity to work side-by-side with them and critical to success. Things like human-centric design and social and behavioral appreciate how AI accelerates organizational goals. sciences to drive responsible business practices. The better the blend of skills, the more sustainable the end result. Those leading the pack ensure employees Experience is another differentiator when it comes to success. Nearly half have ongoing formal training, an understanding of how AI applies to their role, of all Strategic Scalers compared with just a third of respondents in Proof and understand and implement responsible AI. Because AI teams are embedded of Concept report having the necessary experience. The sheer volume throughout the organization, and not cordoned off in special projects, the wider speaks for itself: Strategic Scalers have successfully scaled 114 initiatives on average, compared with just 53 for those in Proof of Concept. Companies achieving success scaling AI initiatives are more likely than their Proof of Concept counterparts to ensure their employees are prepared for the journey: Formal Fully understand Understand & training: how AI applies implement 1.5x to role: responsible AI: 2x 1.7x AI: BUILT TO SCALE 16

Realizing the potential Industrializing for Growth is a dynamic destination. It’s important to note that Industrializing for Growth is a dynamic destination that changes as technology evolves. From our experience, we know of three additional variables that speed companies along their journey to the ultimate destination: A data-driven culture where AI is driving exponential returns. Focus on the ‘I’ in ROI Adopt a digital platform Build trust through mindset to scale Responsible AI The C-suite views AI investment as the cost of doing business. They earmark budget for AI The two main objectives of platforms are Responsible AI entails creating a framework recognizing its criticality for future growth acceleration and extended value. Publishing that ensures the ethical, transparent and and spend without the need to prove ROI in data on a platform once for products to accountable use of technologies in a manner advance to justify the investment. consume through APIs and microservices is consistent with user expectations, more cost effective. It also drives scale by organizational values and societal laws and breaking down silos and democratizing data norms. Responsible AI can guard against the and insights enabling greater collaboration use of biased data or algorithms, ensure that and innovation across the enterprise and automated decisions are justified and broader ecosystem of partners. explainable, and help maintain user trust and individual privacy. AI: BUILT TO SCALE 17

Scaling to It’s not just about SPEED It’s about moving deliberately, in the right direction. new heights of It’s not just about MONEY competitiveness It’s about aligning your investments to the right places with the intention of driving large-scale change. There are reams of information on the “what” of AI. It’s not just about MORE DATA But scaling new heights of competitiveness with AI It’s about investing in your data, deliberately yet pragmatically, to drive the right insights. requires understanding the “how.” And at times eschewing conventional wisdom that continues to It’s not just about a SINGLE LEADER emerge as AI evolves. It’s about building multi-disciplinary teams that bring the right capabilities. AI: BUILT TO SCALE 18

Smooth scaling Scaling the exponential power of AI with digital platforms across the enterprise is a journey. Those that take on the lessons of each path will reach a place where business strategies are seamlessly fused with analytics, leverage a reusable data foundation, and scale through platforms. The result: Industrialized growth through unassailable competitive strength in everything from organizational effectiveness to brand perception and trust. Contact the authors to find out more about how to increase value through scaling AI. AI: BUILT TO SCALE 19

About the More than research US$30 billion US$1 - $5 billion Our research involved 1,500 C-suite 198 323 executives from companies with a minimum revenue of US$1 billion in 12 countries around 510 469 the world across 16 industries, with the aim to uncover the success factors for scaling AI. Our research, and that of our partners in our US$10.1 - $30 billion ecosystem, employs ethical and responsible US$5.1 - $10 billion research methods. Respondents reveal their identities voluntarily, we anonymize all data Revenue (n=1,500) from companies in our data set, and report results in aggregate. We commit to not using the data collected to personally identify the respondents and/or contact the respondents. Titles 12 Countries 16 industries Chief Information Officer (CIO) (441) Brazil (115) Italy (113) Banking & Capital Markets (100) Life Sciences Chief Financial Officer (CFO) (231) Canada (113) Japan (117) Chemicals (100) (Pharma & Biotech) (100) Chief Operating Officer (COO) (215) China (139) Singapore (101) Communications (100) Metals and Mining (100) Chief Digital Officer (136) France (105) Spain (106) Consumer Goods & Services (100) Retail (100) Chief Innovation Officer (114) Germany (116) United Kingdom (116) Energy (Oil & Gas) (100) Software & Platforms (100) Chief Data and/or Analytics Officer (113) India (126) United States (233) Healthcare (Payers) (100) Travel & Transport Chief AI Officer (93) High Tech (100) (Hotels & Passenger) (100) Chief Strategy Officer (59) Industrial Equipment (100) Utilities (100) VP/SVP of AI/ Data/Analytics (98) Insurance (100) AI: BUILT TO SCALE 20

Join the conversation @AccentureStrat www.linkedin.com/company/ accenture-strategy @AccentureAI www.linkedin.com/company/ accentureai/ Contributors Dhruv Jain Lynn LaFiandra Senior Principal, Accenture Strategy Senior Principal, Accenture Research Sandra Reese Chad Vaske Managing Director, Accenture Strategy Senior Manager, Accenture Applied Intelligence AI: BUILT TO SCALE 21

About Accenture About Accenture Applied Intelligence Accenture is a leading global professional services company, Applied Intelligence is Accenture’s approach to scaling AI for our clients. providing a broad range of services and solutions in strategy, We embed AI-powered data, analytics and automation capabilities into consulting, digital, technology and operations. Combining business workflows to accelerate time to value. Our expertise in defining unmatched experience and specialized skills across more than 40 end-to-end strategy, combined with deep data infrastructure industries and all business functions — underpinned by the world’s capabilities, cognitive services and industrialized accelerators help largest delivery network — Accenture works at the intersection of smooth clients’ path to AI adoption, extending human capabilities and business and technology to help clients improve their performance supporting clients in scaling AI responsibly. Recognized as a leader by and create sustainable value for their stakeholders. With 482,000 industry analysts, we collaborate with a powerful global alliance, people serving clients in more than 120 countries, Accenture drives innovation and delivery network to help clients deploy and scale AI innovation to improve the way the world works and lives. within any market and industry. Visit us at www.accenture.com Follow @AccentureAI and visit accenture.com/appliedintelligence About Accenture Strategy About Accenture Research Accenture Strategy combines deep industry expertise, advanced Accenture Research shapes trends and creates data-driven insights analytics capabilities and human-led design methodologies that about the most pressing issues global organizations face. Combining enable clients to act with speed and confidence. By identifying the power of innovative research techniques with a deep clear, actionable paths to accelerate competitive agility, Accenture understanding of our clients’ industries, our team of 300 researchers Strategy helps leaders in the C-suite envision and execute and analysts spans 20 countries and publishes hundreds of reports, strategies that drive growth in the face of digital transformation. articles and points of view every year. Our thought-provoking For more information, follow @AccentureStrat or visit research—supported by proprietary data and partnerships with www.accenture.com/strategy leading organizations, such as MIT and Harvard— guides our innovations and allows us to transform theories and fresh ideas into real-world solutions for our clients. For more information, visit www.accenture.com/research Copyright © 2019 Accenture. All rights reserved. Accenture and its logo are registered trademarks of Accenture.