Agentic AI
,
Artificial Intelligence & Machine Learning
,
Next-Generation Technologies & Secure Development
EXL Study Finds Data, Business Processes Separate Leaders from Laggards
When it comes to artificial intelligence, many business leaders need a reality check.
See Also: AI Agents Introduce a New Insider Threat Model
According to the third annual EXL U.S. Enterprise AI Study, 76% of companies surveyed believe they are ahead of their competitors on AI, but when researchers drilled down into how deeply AI had been embedded across business functions including customer service, marketing, finance and human resources, only 10% were actually AI leaders.
The report segmented respondents into three categories. “Leaders” have fully developed AI capabilities in six to eight functions. “Followers” meaningfully deployed AI in three to five functions, and “Laggards” had done so in two or fewer. The study surveyed 322 C-suite executives and senior decision-makers across the banking and finance, insurance, retail, utilities, life sciences and healthcare payer industries.
“Every company is now using AI in some capacity,” said Anand Logani, executive vice president and chief AI officer at EXL. “But we’re really starting to see leaders distinguish themselves from the pack when it comes to large-scale enterprise integration.”
The data shows that the consequences for overestimating your AI success go beyond having a bruised ego. It shows up in financial results.
Leaders report that AI has trimmed costs by an estimated 26%, boosted revenue by an estimated 27% and improved margins by an estimated 22% in the specific workflows where it has been applied, the survey found. Laggards trailed across all categories.
What separates leaders from those falling behind, Logani said, is a willingness to stop treating AI as an add-on and start redesigning workflows around it from the beginning. Leaders are deploying AI at scale and reshaping their operating models to take advantage of it.
Logani said that understanding why so many companies overestimate their AI maturity requires understanding how the influx of AI into the enterprise tech stack differs from previous technology cycles.
For example, when companies began moving to cloud, business leaders signed off on the investment but didn’t get involved on a granular level. But with AI, boards and the C-suite are taking an intense interest in how and where AI is being deployed, and they know more about AI than they did about cloud, ERP systems or data warehousing.
“Because they are plugged in, there’s a lot of education flowing up to the C-level suite,” Logani said. “So, relatively you’re more informed than the past wave.” But being well-briefed about AI and having deployed it on a meaningful scale are not the same thing.
Early success compounds the problem. An organization that has gotten a pilot into production has real evidence that AI works. Companies then compare themselves to the firms they hear about struggling in the media, and they feel ahead, Logani said.
The other issue, Logani said, is the lack of benchmarks. “The benchmark of ‘good’ is just relative to where you think you are and where you thought you would be,” he said. “If you’re meeting those expectations, you feel you’re ahead, but what best-in-class looks like hasn’t yet been fully established.”
That gap between confidence and maturity becomes especially visible in the way companies manage data. Even organizations with promising AI pilots can struggle to scale them if the underlying data is fragmented, inaccessible or poorly governed. Seven in 10 respondents said data was a challenge in effective AI use.
Data privacy and security was cited as an obstacle by 24% of respondents, and 31% said siloed data residing across multiple sources was a problem. Also, 58% said they lacked the skills to effectively use AI to capitalize on their data, and 61% reported they can’t consistently and quickly access data to support timely AI-enabled decisions.
The gap between leaders and laggards in data management was significant. Data is accessible enterprise-wide for 44% of leaders but for just 17% of laggards. Meanwhile 91% of leaders said they use best practices or have leading-edge data management practices, while only 61% of laggards said the same.
For CIOs looking to shore up their data strategies, Logani said not to start with a sweeping data consolidation agenda. “Don’t undertake a five-year data consolidation strategy. That era is completely gone,” he said. Instead, he recommends choosing high-impact use cases and working backwards from the desired outcome to determine what data strategy, architecture, context layer and semantic layer those use cases require.
“Galvanize yourself to those high-impact areas and work backwards,” he said. “You have made the choices of the strategy, you have made the choices of architecture, you have made the choices of your context semantic layer. Then you can keep building and layering on that architecture based on high-impact cases that you will implement.”
Rethinking business processes is emblematic of the way leaders have transformed their operating models. The survey found that 44% of leaders said they have completely redesigned their enterprise operating models to accommodate AI. For laggards, only 23% had done the same.
This distinction matters, Logani said, because most organizations making “significant changes” are not merely shifting tasks to AI while leaving workflows intact, they’re redesigning processes.
“If AI were embedded from the start, how would this workflow, this team, this decision look different?” Logani said. “That is still very rarely adopted at scale, but clearly people do realize that reimagination and operating model transformation is a core pillar of AI.”