Executives who fail to embrace AI may soon fall behind. Businesses with AI-literate leaders will outperform competitors by 20% financially by 2027, according to Gartner’s data and analytics predictions for 2025.
“Unlocking AI’s full business potential requires building executive AI literacy,” the report stated. “They must be educated on AI opportunities, risks and costs to make effective, future-ready decisions on AI investments that accelerate organizational outcomes.”
Additional predictions ranged from the impact of AI on data-driven decision making and board-level decisions to the utility and dangers of synthetic data.
How AI agents will be used in business decisions
Gartner predicts AI agents for decision intelligence will be used in 50% of business decisions by 2027. In decision intelligence, leaders leverage a combination of data, analytics, and AI to drive business growth.
Some business leaders deploy AI agents to retrieve and analyze complex data sources. However, while generative AI is improving at analyzing unstructured data, Gartner acknowledges that AI should not be the sole decision-maker.
“AI agents for decision intelligence aren’t a panacea, nor are they infallible,” said Carlie Idoine, vice president analyst at Gartner, speaking at the Gartner Data & Analytics Summit in Sydney on June 17. “They must be used collectively with effective governance and risk management. Human decisions still require proper knowledge, as well as data and AI literacy.”
To avoid pitfalls, business leaders should focus on active metadata in data management and prioritize semantics in AI-ready data.
Prioritizing semantics in particular can reduce hallucinations and decrease the number of tokens used, Gartner said; this, in turn, reduces costs. Focusing on semantics is likely to improve GenAI model accuracy by up to 80% and reduce costs by up to 60%, according to Gartner.
AI can also be used in board-level decisions. Gartner predicts that 10% of global boards will employ AI guidance to challenge executive decisions material to their business by 2029. The advisory firm also advised boards to establish clear policies around oversight, responsibility, and regulatory compliance.
Using synthetic data in decision-making can pose problems
AI is sometimes used to analyse synthetic data, which generative AI models create by extrapolating from real information.
Synthetic data is often used in AI tasks to train models from diverse datasets, but it can also pose challenges. Synthetic data may be inaccurate or difficult to fold into existing data pipelines and systems. Gartner predicts that 60% of data and analytics leaders will encounter critical failures in managing synthetic data by 2027.
Experiential upskilling for executives is key
To effectively loop AI into decision-making, executives must understand the associated benefits, risks, and costs. Gartner recommends that data and analytics leaders offer experiential upskilling for executives — such as building domain-specific prototypes — to make AI more tangible and drive smarter, more targeted investments.
Read more about the intersection of AI and business intelligence.
