*SAS floats Trustworthiness Index to measure business impact

*As businesses still search for AI profit

By Prince Osuagwu, Hi-Tech Editor

As businesses increasingly turn to artificial intelligence, AI, to drive efficiency, innovation and faster decision-making, trust has emerged as one of the biggest issues threatening the technology’s ability to deliver greater business value.

A new report by SAS, a global leader in data and AI software, with insights from IDC, shows that while organisations are rapidly deploying increasingly autonomous AI systems, the governance, data quality and oversight needed to make such systems trustworthy are not keeping pace.
The findings are contained in the 2026 Data and AI Impact Report: The New Economics of Trust, which examines what organisations need to deploy increasingly autonomous AI responsibly, confidently and at scale.

To measure this, SAS introduced a Trustworthiness Index, based on five dimensions: Data Quality and Governance; Model Governance and Oversight; Explainability and Fairness; Responsible AI Policy; and Audit and Accountability.
Organisations scoring 80 or higher are classified as trustworthiness leaders.

Interestingly, the research found that organisations with stronger governance, data quality and auditability practices — although still a relatively small segment — consistently outperformed their peers, reporting at least twice the return on investment (ROI) from their AI deployments.
By contrast, fewer than one in 20 organisations classified as AI trustworthiness laggards reported similar levels of ROI.
“When AI works, it’s incredibly impactful,” said Bryan Harris, Chief Technology Officer at SAS.
However, he noted that state-of-the-art AI agents can have error rates exceeding 25 per cent on complex tasks, a situation he said was unacceptable in high-stakes decision-making.
According to Harris, organisations seeking accuracy and repeatability must embed domain expertise into AI-agent workflows while keeping humans at the centre of governance and oversight.
Chris Marshall, Vice President at IDC, said the growing autonomy of AI was creating a new trust challenge for businesses.
“As AI becomes more autonomous, organisations face a new challenge: maintaining confidence in systems people don’t fully understand,” he said.
He added that stronger oversight, explainability, accountability and data foundations were becoming prerequisites for scaling AI successfully.
When AI starts acting, trust drops
One of the striking findings of the report is that trust falls as AI moves from simply generating information to taking action.
While 66 per cent of respondents said they trusted agentic AI, the figure was 10 percentage points lower than trust in generative AI.

At the same time, about 89 per cent of respondents said their AI agents already play a role in decision-making, ranging from informing human-approved decisions to acting autonomously.

Yet organisations with trustworthy AI practices were found to be 15 times more likely to report strong or high ROI — 62 per cent compared with just four per cent among less trustworthy organisations.

The implication is that trust is no longer simply an ethical or compliance issue. It is increasingly becoming a business issue.
Why users override AI
The research also identified explainability as the biggest barrier to trust.
“Lack of explanation” was cited as the top reason users override AI decisions, accounting for 34.5 per cent of overrides.
Explainability and Fairness also recorded the least improvement among the five dimensions measured by the Trustworthiness Index.

For organisations in Nigeria, where businesses are increasingly exploring AI for financial services, customer engagement, cybersecurity, healthcare, logistics and other operations, the issue could become even more important as AI systems move beyond producing recommendations to taking actions.

The data problem

Another major challenge identified by the report is the weakness of the data foundations supporting AI.
Only 17.5 per cent of organisations operate data infrastructure at the highest maturity level, according to the study.
Many organisations are deploying AI on outdated or underdeveloped data infrastructure, making it difficult to provide the data lineage and validation required for increasingly autonomous systems.
Without reliable data, the report suggests, businesses may struggle not only to explain how AI reaches decisions but also to govern the technology effectively and extract the expected value from their investments.

The findings point to a changing phase in the AI race.
As businesses move from systems that generate answers to systems that can make decisions and take action, the report suggests that the ability to trust those systems could increasingly determine how much value companies ultimately get from their AI investments.