AI agents are only as good as the data they're given, and that's a big issue for businesses

zf L/Moment/Getty Images

Follow ZDNET: Add us as a preferred source on Google.

ZDNET's key takeaways

• 63% of business leaders describe their organizations as very data-driven, up 10% from 2023. 
• Only one in two business leaders is confident about the ability to deliver timely business insights.
• The most valuable insights for organizations are currently trapped in unstructured data. 

Business leaders understand the value of data. Sixty-three percent of today's business leaders describe their organizations as very data-driven, up 10% from 53% in 2023, according to Salesforce's State of Data and Analytics Report based on a survey of 3,800 data and analytics leaders and 3,852 cross-functional business leaders worldwide. That said, nearly two-thirds (63%) of technical leaders acknowledge that their companies struggle to drive business priorities with data.

The rapid emergence of AI agents has created a palpable sense of urgency for companies across all industries and geographies, as they seek ways to reinvent themselves as 'agentic enterprises' to accelerate business growth. Salesforce's research revealed that business, data, and analytics leaders are grappling with how to overhaul their data infrastructure, management, and governance. 

Also: Enterprises are not prepared for a world of malicious AI agents

This shift is critical for both success -- or in some cases, survival -- as their companies prepare for an agentic future. This future requires the democratization of data, AI, and analytics, making these resources accessible to everyone within the organization. 

Here are the four key findings from the Salesforce report: 

1. AI accelerates a new data paradigm: The vast majority of companies now use at least one form of AI in their day-to-day workflows. As these revolutionary technologies take hold, business and technical leaders are questioning how prepared their underlying foundations and cultures are, especially as data volume and complexity increase. To reiterate, 63% of business leaders describe their organizations as very data-driven.
2. Limited data confidence thwarts activation, decisions, and action: As organizations become more data-driven, many business leaders feel lost in the slow, technical processes of generating analytic insights. What's more, many aren't sure the data they rely on is accurate in the first place. Fifty percent of business leaders aren't sure they can generate and deliver timely insights.
3. Building the data foundation for analytics and AI: On the technical side, data and analytics leaders are pressured from increased line-of-business demand for data-driven capabilities and executive demands for agentic innovation. But poor data management practices, including integration and harmonization, plus the dominance of unstructured formats, pose formidable challenges. Seventy percent of data and analytics leaders believe the most valuable insights for their organizations are trapped in unstructured data.
4. Governing and safeguarding the agentic enterprise: The rise of AI is exposing long-standing shortcomings in data security, compliance, and governance measures. Only 43% of data and analytics leaders have established formal data governance frameworks and policies, and 88% believe AI demands new approaches.

Here are my key takeaways from the top two insights: AI accelerates a new data paradigm, and limited data confidence limits decisions and actions. 

AI accelerates a new data paradigm

Ninety percent of business leaders believe their careers depend on being data-fluent. Meanwhile, 86% believe their careers depend on being data-driven, and I think the other 14% will be looking for new jobs soon. 

There is, however, a disconnect in business -- although businesses report using data more, technical leaders have reservations, with nearly two-thirds (63%) agreeing their companies struggle to drive business priorities with data. Data and analytics leaders estimate that 26% of their organizations' data is "untrustworthy." And 42% of business leaders said their data strategies don't fully align with business objectives.

Also: The best AI agents are terrible freelancers - for now

All AI projects are data projects, and nearly all organizations agree that the rise of AI makes data more critical. Ninety-three percent of organizations have at least one instance of AI in their technology stacks, according to Salesforce's latest State of IT survey. The technology's rapid evolution, including the advent of agents, is putting pressure on data and analytics leaders to ramp up capabilities quickly. Leaders see AI as a forcing function to improve their overall data literacy and culture, with 91% of business leaders believing that the rise of AI makes it more important to be data-driven.

So, how are businesses changing their investment thesis to support AI programs, including the adoption of agentic AI? The good news is that there are boundless opportunities to improve efficiency, innovation, and productivity using AI agents. The State of IT research from Salesforce found that 84% of CIOs believe AI will be as significant to their businesses as the internet. 

Also: Even the best AI agents are thwarted by this protocol - what can be done

Additionally, 84% of data and analytics leaders agree AI's outputs are only as good as its data inputs. Another key finding is that CIOs spent four times as much of their budget on data infrastructure as on AI. The reason for that investment may be that 80% to 90% of enterprise data is estimated to be unstructured, and 70% of data and analytics leaders believe the most valuable insights for their organizations are trapped in unstructured data.

There is an ocean of data, and yet we are thirsty for insights. Data and analytics leaders estimate their organizations' data volumes grow 30% annually, up from 23% in 2023. The average number of data environments across the enterprise is: 

• 26 spreadsheet applications
• 21 cloud storage services
• 21 operational databases
• 17 data warehouses
• 16 data lakes
• 15 customer data platforms

The average enterprise uses 897 applications, and only 29% are connected. Business leaders generally don't fully trust their data, citing persistent issues such as accuracy, reliability, and relevance. More than half (54%) of leaders aren't entirely confident that the data they need is accessible in the first place. Data and analytics leaders estimate 19% of their companies' data is trapped.

Also: Want better ChatGPT responses? Try this surprising trick, researchers say

The top data priorities are: building AI capabilities, providing real-time data access, improving company-wide data fluency, improving data quality, and strengthening security and compliance. The top data challenges are: lack of real-time data, lack of data harmonization, security threats, ensuring data accuracy and quality, and siloed or trapped data. 

Real-time data rose dramatically as the top data challenge, surpassing perennial pain points like harmonization, security threats, and overall accuracy and quality. And as AI underscores the need to derive value from unstructured data, trapped and siloed sources round out the top five data hurdles, skyrocketing from last of all concerns two years ago.

Limited data confidence

Regardless of whether it's powering an AI prediction, an agent-driven customer interaction, or a report highlighting key metrics, data is only good if it's grounded in the business context that contributes to "trustworthy data." Ninety-three percent of business leaders agree insights are only relevant if they're grounded in the business context. Here are the top factors preventing data-driven organizations:

1. Incomplete, out-of-date, or poor-quality data.
2. Lack of tools to access, analyze, and interpret data.
3. Lack of expertise and training on how to access, analyze, and interpret data.
4. Takes too long to get insights.
5. Lack of access to the required data.

The challenges noted above prevent employees from acting on relevant insights promptly, with 49% of data and analytics leaders saying their companies occasionally or frequently draw incorrect conclusions from data that misses or misunderstands business context.

Greater adoption of AI is accelerating how businesses access and act on data. A vast majority of analytics and data leaders (91%) said technical queries limit analytics use at scale, and 92% cited a lack of data fluency among staff. AI is, therefore, increasingly used to augment analytics processes, with 64% of business leaders using AI to find, analyze, and interpret data, while only 54% get help from a technical resource. 

Greater adoption of AI means buyers have higher expectations from AI solutions. Analytics solutions buyers prioritize AI and real-time data, including AI-driven actions -- 88% of data and analytics leaders said advances in AI are changing how they evaluate analytics software and implementations. Analytics and data leaders are looking for real-time data, AI-assisted workflows, AI-driven actions, composable analytics, and insights at scale. The report found that 94% of business leaders said they'd perform better with direct data access in the programs/apps where they work the most.

Also: AI is more likely to transform your job than replace it, Indeed finds

Analytics and data leaders embrace AI agents to assist and drive outcomes for their stakeholders. AI agents that can understand, respond to, and take actions based on user inquiries in natural language hold particular appeal and potential. Specifically, agentic analytics makes data consumption and interaction highly intuitive and conversational. Leaders want to have conversations with their data platforms. As many as 63% of data and analytics leaders said translating business questions into technical queries is prone to error, and 93% of business leaders said they'd perform better if they could ask data questions with natural language.

To learn more about the State of Data and Analytics report, you can visit here.