top of page

AI Requires Time, Scale, and Skill

Firms will invest more in AI after the pandemic, but delivering ROI will take skill, scale, and time.

ESI ThoughtLab study of 1,200 organizations reveals that generating ROI on AI is still a work in progress that requires a focus on strategic change.

Two-thirds of senior executives across industries—and nearly nine out of ten leaders from the world’s largest enterprises—believe that artificial intelligence (AI) is important for the future of their businesses and will be upping their investments in the post-pandemic era.

Yet their companies are now seeing an average ROI of only 1.3%, and 40% of AI projects are not yet profitable according to Driving ROI through AI, a just-released research study conducted by ESIThoughtLab and a coalition of AI leaders, including Appen, Cognizant, Cortex, Dataiku, DataRobot, Deloitte, and Publicis Sapient.

The reason for this paradox, according to the research, is that AI initiatives require time, expertise, and scale to deliver on their promise of high returns. With the pandemic speeding up the need for quick data-driven decision-making, companies should act now to develop the skills, platforms, and processes that can enable them to achieve the full strategic, operational, and financial benefits from AI.

ESI ThoughtLab economists benchmarked the AI practices, performance results, and three-year plans of 1,200 companies in 12 industries and 15 countries, which together have combined revenue of $15.5 trillion (or about $12.9 billion per firm).

Conducted during the COVID-19 outbreak, the study reveals the value that AI can bring in a socially distancing, the digital-first world—including access to time-critical data, event-driven forecasts, personalized digital experiences, flexible work processes, rapid decision-making, tighter cybersecurity, and greater cost efficiencies.

Executives Should Not Expect Fast Results

The research shows that delivering ROI on AI can be elusive for the uninitiated and slow going even for experienced organizations. Those in earlier stages of AI adoption often see flat results. It is not until they scale AI more widely across their enterprises and become leaders that the ROI rises to 4.3%.

With frequently high upfront costs in data preparation, technology adoption, and people development, it takes an average of 17 months for a firm to reach break-even and months more to generate significant returns.

Most companies, even leaders, are still relatively early in their AI journey. Only about one-quarter of AI projects are now in widespread deployment among AI leaders. Many AI projects are still in pilot or early deployment stages. However, firms are planning to boost their AI investments by an average of 8.3% annually over the next three years, bringing their annual AI spend from $38 million currently (or 0.75% of revenue) to over $48 million.

The ROI of AI Comes From Strategic Change

As companies progress in AI use, they often shift their focus from automating internal employee and customer processes to delivering on strategic goals. For example, 31% of AI leaders report increased revenue, 22% greater market share, 22% new products and services, 21% faster time-to-market, 21% global expansion, 19% creation of new business models, and 14% higher shareholder value. In fact, the AI-enabled functions showing the highest returns are all fundamental to rethinking business strategies for a digital-first world: strategic planning, supply chain management, product development, and distribution and logistics.

The study found that automakers are at the forefront of AI excellence, as they rush to use AI to deliver on every part of their business strategy, from upgrading production processes and improving safety features to developing self-driving cars. Of the 12 industries benchmarked in the study, automotive employs the largest AI teams (557 people on average vs. 370 for all industries) and has the largest AI budgets ($59.4 million on average vs. $38.3 for all industries). With the government actively supporting AI under its Society 5.0 program, Japanese companies lead the pack in AI adoption. Unlike in the US, where AI is viewed often as a threat to jobs, firms in Japan tend to see AI a way to fill the employment gap caused by an aging population and stringent immigration laws.

Driving High ROI From AI

To drive AI performance, executives should consider these best practices uncovered by the research:

  1. Begin with pilots, then scale AI applications across the enterprise. Companies starting out should work closely with business teams to identify use cases and demonstrate AI’s worth through pilots. But the true value of AI can materialize only with widescale deployment when firms can offset their upfront costs with substantial business gains.

  2. Lay a firm foundation. Organizations should have the proper IT and data management system in place; have a secure and sufficient budget; work through the data security, privacy, and ethical risks of AI; develop a clear vision and plan that takes into account AI-driven strategic transformation; obtain senior management support, and have a robust ecosystem of partners and suppliers.

  3. Get your data right. Nine out of ten AI leaders are advanced in data management. But ensuring your data is in good shape is not enough; organizations should bring in a diverse set of data, such as psychographic, geospatial, and real-time data. The study found that combining different types of data can create a multiplier effect on AI returns.

  4. Solve the human side of the equation. AI is as much about people as technology. AI leaders spend 27% of their AI budget on developing and hiring people, almost twice the percentage that AI beginners spend. They are also more apt to appoint specialists, such as Chief AI and Data Officers, to lead their AI initiatives. They outsource less and build internal teams more.

  5. Adopt a culture of collaboration and learning. About 85% of companies that generate large AI returns work to ensure close collaboration between AI experts and business teams. AI leaders are better at providing non-data-scientists with AI skills. They also decentralize AI authority to help ensure that AI responsibility and expertise are distributed across their organizations.

“As the pandemic propels businesses into a digital-first world, AI will become a key driver of corporate growth and competitiveness. But building proficiency in AI is not easy,” said Lou Celi, ESI ThoughtLab CEO and program director for Driving ROI through AI. “AI is not a magic bullet. It can fail to deliver results if the wrong business case is selected, the data is prepared incorrectly, or the model is not built for scale.”


bottom of page