Trustworthy AI practices drive high ROI and business success

SAS, working alongside research analysts from IDC, recently published the second annual Data and AI Impact Report: The New Economics of Trust to explore this phenomenon.

The research team surveyed 2,699 decision-makers across 28 distinct countries, focusing specifically on highly regulated sectors like banking, insurance, life sciences, and the public sector. These comprehensive findings confirm that ethical and transparent technological practices function as primary economic drivers rather than mere regulatory compliance exercises.

The Widening Profitability Gap

Financial metrics reveal a stark contrast between enterprises prioritising governance and those neglecting these crucial frameworks. The comprehensive report highlights a rapidly expanding return on investment (ROI) divide across the current global market.

Several key statistics illustrate this disparity vividly:

• Enterprises actively investing in trustworthy AI measures are 15 times more likely to report strong or high ROI on their specific projects.

• Exactly 62% of these mature organisations achieve strong returns, compared to just 4% among less prepared industry peers.

• Market leaders realise 1.85 times greater gains across 13 distinct business outcomes, spanning vital areas like revenue growth, cost savings, and customer experience enhancements.

• Recognising this distinct advantage, 85% of leaders are increasing their investment in trustworthy practices by more than 10% this calendar year.

Organisations capturing the most immense value are not necessarily acquiring vastly different software tools. They are instead managing their technological assets through a fundamentally different, trust-centric lens.

The Explainability Crisis

Sophisticated technical capabilities mean very little if human operators refuse to trust the final output. The extensive study uncovered a significant operational friction point regarding workforce adoption. A staggering 97.2% of users routinely override AI-generated recommendations in at least some specific instances.

The primary catalyst for human intervention relates directly to systemic explainability. Employees are consistently hesitant to rely on automated systems that cannot articulate the rationale behind a specific decision or recommendation. Artificial intelligence that lacks the capacity to explain its own internal logic becomes a major business liability, especially within high-stakes environments like healthcare diagnostics or financial risk assessment.

Bridging the widening gap between human expertise and machine intelligence is therefore absolutely critical. Bryan Harris, CTO at SAS, addressed this operational necessity very clearly.

"When AI works, it's incredibly impactful," Harris stated. He added that "in order to achieve accuracy and repeatability, organizations must embed domain expertise into agentic workflows, while keeping people at the center of governance and oversight." Harris concluded that "organizations that do this successfully will close the trust gap and gain a competitive advantage in the market with AI."

Core Pillars of Reliable Systems

Researchers evaluated each surveyed enterprise across five strict criteria to benchmark maturity levels accurately. Systems must excel across multiple distinct dimensions to be considered truly reliable and secure by the workforce.

The five core pillars include:

• Data quality and governance: Guaranteeing the foundational information feeding the complex models remains accurate, unbiased, and secure.

• Model governance and oversight: Implementing continuous human-in-the-loop supervision over all active algorithms.

• Explainability and fairness: Demystifying complex automated decision-making processes so end users understand the underlying logic completely.

• Responsible AI policy: Creating and enforcing comprehensive ethical guidelines that dictate enterprise usage.

• Audit and accountability: Establishing clear, predetermined chains of responsibility to ensure strict regulatory compliance at all times.

The Data Foundation Imperative

Trustworthy machine learning cannot function without a highly optimised data infrastructure. Most organisations currently lack the robust data foundation required to make their initiatives trustworthy and ultimately profitable. The research indicates that only a comparatively small market segment currently possesses the necessary maturity to scale these technologies successfully.

Enterprises operating with an optimised data foundation hold a massive competitive advantage. They are four times more likely to expect strong financial returns from their automated projects. These same progressive companies are six times more likely to mandate the strict data quality controls necessary to build lasting trust across the workforce.

Robust oversight and explainability are no longer optional additions to a corporate technology stack. Strong data foundations and accountability measures function as absolute prerequisites for scaling modern automated solutions. Companies that recognise and successfully implement these vital safeguards will undoubtedly secure the greatest long-term commercial success in the coming decade.