AI Technologies , Generative AI , Standards & Best Practices
Beyond the Hype: Why AI's Payoff Will Take Time
94% CIOs Embrace AI But Only 14% Expect to Be AI-Ready by 2025, Says MIT ReportAI's promise to transform business operations has been met with reality checks according to the recent MIT Technology Review report titled "CIO Vision 2025: Bridging the Gap Between BI and AI." Despite the widespread enthusiasm related to AI adoption, many businesses are still struggling to implement AI effectively.
See Also: Business Rewards vs. Security Risks of Generative AI: Executive Panel
The report, sponsored by Databricks, showed that while 94% of CIOs are embracing AI, more than half expect it to be widely adopted by 2025. Despite its potential to enhance customer experiences and streamline operations, companies are encountering major obstacles in realizing their AI goals.
Successful AI adoption hinges on unified data platforms and well-defined strategies, according to Stephanie Woerner, a research scientist at MIT's Center for Information Systems Research. "CIOs must navigate the gap between ambition and execution. Focusing on data quality, closing skills gap and establishing clear frameworks for AI ethics are critical steps for CIOs," she said.
The report collected insights from 600 global CIOs, CDOs and CTOs spanning 14 industries. Companies such as Procter & Gamble, Johnson & Johnson, Cummins, CNH Industrial, Walgreens Boots Alliance, S&P Global, Marks & Spencer, Tokio Marine, Virgin Australia and Freshworks have already integrated AI into their operations to varying degrees. Despite this, only 14% of these organizations are aiming to be fully AI driven by 2025 - a goal that is proving increasingly difficult to achieve.
Data: The Biggest Challenge
The report showed that data-related challenges are the primary obstacle, with 72% of respondents identifying this issue. Ensuring proper data management and quality is essential for the success of AI initiatives.
CIOs and technology leaders recognize the need to organize their data to achieve their AI objectives, with 78% of enterprise tech leaders planning to prioritize scaling AI and ML to generate business value within the next three years. But even though these organizations have potential AI use cases, many are discovering that the anticipated benefits are not materializing as quickly as expected. For most companies, scaling AI projects beyond small pilot programs is proving to be difficult.
Data quality remains a significant challenge for AI, according to Mike Maresca, global CTO at Walgreens Boots Alliance. "We have the right platform, tools and governance in place, but maintaining high data quality while enhancing algorithms is critical as we scale," he said (see Image 1).
AI Leaders Face Hurdles
Leading companies such as P&G, Johnson & Johnson and Cummins are advancing in AI adoption, but they continue to encounter significant data challenges. Cummins, a power engineering firm, began using AI five years ago to enhance customer services but initially struggled to generate new revenue streams, as customers were unwilling to pay extra for features they considered part of the core product. This prompted Cummins to pivot its AI strategy toward cost-saving initiatives, including predictive maintenance. "This changed how we viewed AI and the data generated by our engines," said Sherry Aholm, chief digital officer at Cummins.
S&P Global's CIO Swamy Kocherlakota emphasized the need for scalable data processing. "We're dedicating significant time to understanding how to apply AI/ML and NLP at scale." Retailers such as Marks & Spencer are also enhancing their data infrastructure to support larger AI initiatives, with a focus on data quality, searchability and governance (see Image 2).
Heavy Investment, Uncertain Returns
Investments in AI adoption are rising despite these challenges. The report indicated that 78% of respondents prioritized scaling AI and ML to generate business value.
On the one hand, many organizations, including Freshworks, are hesitant to broaden their AI initiatives without clear return on investment. For Johnson & Johnson, on the other hand, AI investments have yielded significant gains in productivity and decision-making.
"We've seen increased productivity, better risk mitigation from human error, and faster and more insight-driven decision-making," said Rowena Yeo, CTO at Johnson & Johnson. "AI is accelerating clinical trials, which directly impacts revenue," she said.
AI at Scale: Not Yet There
The report showed that many organizations struggled with scaling AI beyond pilot projects. While many companies are experimenting with AI, transitioning from trials to comprehensive adoption has proven difficult. P&G and Tokio Marine are both striving to automate their AI processes, yet a notable gap remains between AI leaders and those struggling to scale (see Image 3).
Tokio Marine - one of Japan's oldest insurance companies - is using AI to streamline its claims process. But in their journey to become fully AI-driven, the company is grappling with integrating legacy systems and ensuring data readiness. Although some of Tokio Marine's initiatives - including using computer vision to assess auto insurance claims - show promise, scaling these solutions across the organization will necessitate both technological and cultural shifts.
"Current development efforts focus on analyzing data from in-car drive recorders to monitor driver actions and behaviors. Additionally, enhancing fraud detection with AI is a top priority for the company," said Masashi Namatame, group chief digital officer at Tokio Marine.
P&G is using AI to enhance product development and consumer research, but it is facing challenges in automating its entire AI process. To successfully scale AI across its extensive operations, the company must prioritize automation and democratization, empowering non-data scientists to manage AI models. Without these initiatives, P&G's progress could be hindered.
"We aim to create more AI use cases by automating the entire AI life cycle," said Vittorio Cretella, CIO of P&G. The company also seeks to democratize AI by allowing more employees to configure algorithms using user-friendly platforms, reducing dependence on data scientists.
The Road Ahead
The path to AI adoption presents both promise and challenges. While AI has significant potential, companies must overcome data issues, resolve scaling hurdles and ensure strategic clarity to unlock its benefits. Those that navigate these obstacles will become AI leaders, while others may struggle with the gap between hype and reality. Although excitement related to generative AI may give CIOs some breathing room, a KPMG survey of 1,300 corporate leaders indicated that most anticipate a return on investment within three to five years, not right away.