Study: 84% of Technical Leaders Need Data Overhaul for AI Strategies to Succeed

Seventy-six percent of business leaders say they're under growing pressure to drive business value with data. Yet Salesforce's new State of Data and Analytics report reveals their biggest hurdle is still incomplete, out-of-date, or poor-quality data. The chasm between businesses' data demands and their data realities becomes more problematic in the agentic AI era. While business leaders are eager to use AI for insights and productivity, their technical counterparts worry a new approach to data and analytics is needed. In fact, 84% of data and analytics leaders say their data strategies need a complete overhaul before their AI ambitions can succeed.

To close the gap, savvy technical leaders are focusing on the fundamentals: timely, context-rich data; stronger governance; and zero copy architectures that unlock trapped, distributed data regardless of where it resides. On their journey to becoming agentic enterprises, they're also embracing emerging solutions, like agentic analytics, that bring reliable insights into the flow of work.

The lessons learned from earlier waves of AI adoption provide a blueprint for companies to become agentic enterprises, where human employees and intelligent AI agents work together.

"The lessons learned from earlier waves of AI adoption provide a blueprint for companies to become agentic enterprises, where human employees and intelligent AI agents work together," explained Michael Andrew, Chief Data Officer at Salesforce. "Trusted, unified, and contextual data is the key that unlocks everything. For organizations ready to execute at scale, this is the moment to shore up data foundations to confidently scale AI to its full potential to deliver real value and ROI."

Existing data foundations strain to support business ambitions.
Nearly two-thirds of business leaders (63%) describe their organizations as data-driven - up 10 points from 2023. Yet just as many (63%) data and analytics leaders say their companies struggle to drive business priorities with data, exposing a gap between data maturity perceptions and reality.

• Less than half (49%) of business leaders say they can reliably generate timely insights.
• Nearly half (49%) of data and analytics leaders say their companies occasionally or even frequently draw incorrect conclusions from data with poor business context.
• Incomplete, out-of-date, or poor-quality data remains the number one factor preventing organizations from being truly "data-driven."

Poor data derails path to becoming an agentic enterprise.
AI has quickly become the top data priority - and the biggest stress test for existing data foundations. In 2023's State of Data and Analytics report, expanding AI capabilities was among the top three data priorities; today, it's the uncontested number one.

• As a result, 67% of data and analytics leaders feel pressure to implement AI quickly.
• Yet 42% lack full confidence in the accuracy and relevance of their AI outputs, likely because of the disconnected, out-of-date data it draws from.
• While 84% of data and analytics leaders theoretically agree that AI's outputs are only as good as its data inputs, their reality is a bit more complicated. Data and analytics leaders estimate over a quarter (26%) of their organizational data is untrustworthy.

Businesses are feeling the consequences of training AI on faulty data foundations.

• 89% of data and analytics leaders with AI in production say they've experienced inaccurate or misleading AI outputs.
• More than half of data and analytics leaders (55%) at companies training or fine-tuning their own models report they've wasted significant resources doing so with bad data.

89% of data and analytics leaders believe a strong data foundation is the most critical factor for successful AI.

"Agentic AI isn't the next technology - it's the next revolution. AI agents handle routine tasks so humans can focus on creativity, relationships, and impact," announced Salesforce CEO Marc Benioff in his Dreamforce keynote. However, he warned that "to truly get the most value and context from AI models, you've got to get your data right. You have to get to more integrated solutions. You have to get the priorities right. You have to get the governance right."

Even high-quality data is useless if it's trapped.
Nearly 9 in 10 data and analytics leaders believe unified data is key to meeting customer expectations. At the same time, trapped data remains an ever-growing challenge, skyrocketing from last on their list of challenges two years ago to a top five hurdle today. The issue is exacerbated by application sprawl; the average enterprise uses 897 applications, and only 29% are connected. This severe fragmentation scatters data across silos, making it difficult or impossible to access. As a result:

• Data and analytics leaders estimate that 19% of their company's data is siloed, inaccessible, or otherwise unusable.
• More concerning, 70% of data and analytics leaders believe their most valuable business insights reside within this inaccessible 19%.
• The ramifications are widespread, with over 8 in 10 data and analytics leaders citing reduced AI capabilities, obscured customer views, reduced personalization, and missed revenue opportunities as a result.

To meet business demands, technical leaders revisit how data is accessed, used, and secured.

• To mitigate trapped data challenges, 56% of organizations are adopting zero copy data integration, an approach that makes it possible to access data that is sitting in multiple different databases at the same time without having to move, copy, or reformat anything.
  • These changes are paying off. Companies using zero copy are 25% more likely to deliver superior customer experiences and 34% more likely to succeed with AI initiatives than those without zero copy.
• Natural language interfaces, like agentic analytics, can solve for data literacy and access bottlenecks.
  • 63% of data and analytics leaders say translating business questions into technical queries is prone to error.
  • 93% of business leaders say they'd perform better if they could ask data questions with natural language.
• Governance and security protocol updates are needed to address increasingly complex data demands.
  • Only 43% of data and analytics leaders have established formal data governance frameworks and policies.
  • 88% of data and analytics leaders agree that AI demands entirely new approaches to governance and security.

"As companies move towards becoming agentic enterprises, true transformation happens when data and AI move in lockstep, " said Andrew. "Strong data foundations give AI the context it needs, and AI, in turn, helps leaders unlock the full potential of their data. The organizations treating data and AI as an integrated strategy are the ones who will successfully move from pilots to execution to see AI deliver significant impact."

Go Deeper:

Methodology:

Data in this report are from two double-anonymous surveys conducted from June 27 through August 13, 2025. The first survey generated 3,800 responses from analytics and IT decision makers from 18 different countries across North America, Latin America, Asia-Pacific, and Europe. The second survey generated 3,852 responses from line-of-business leaders from the same countries. More details can be found in the report. Cultural bias impacts country-level survey results.