28 August 2026
Trustworthy AI: How to move from ambition to ROI
Cindy Turner
Insights Editor (in consultation with Kristi Boyd, Trustworthy AI Specialist), SAS
While wildly popular in everyday life (and sometimes contentious), artificial intelligence (AI) capabilities have become embedded in everything from writing software code to analyzing policies. The ubiquitous nature of the technology leaves business and government leaders scrambling to answer questions, like:
- Can I (and should I) trust AI?
- Which AI applications should I adopt first, and are they secure?
- What effect will AI have on my business, employees, customers and community?
- Will my AI tools deliver reliable results, or will they create unintended consequences?
- When should I expect to see a return on my AI investments?
Your ability to ask and respond to these big questions can make or break the long-term value your organization achieves from AI systems.
Research backs up this assertion.
According to a new SAS report developed with research insights from IDC, organizations that invest in responsible AI and strong governance see higher ROI – while even the most advanced organizations underperform without it.
What is trustworthy AI?
Trustworthy AI is artificial intelligence technology that is designed, developed and deployed with human-centricity and other essential principles in mind. AI systems that are worthy of trust should incorporate six core principles:
- Human-centricity – promoting human well-being, human agency and equity.
- Accountability – recognizing potential harms before they happen and acting proactively.
- Transparency – openly communicating and providing documentation of an AI system’s intended uses, risks, and how its decisions are made and monitored.
- Inclusivity – ensuring that AI reflects diverse populations and that it works for everyone.
- Robustness – operating reliably and safely, while managing risks
- Privacy and security – protecting the use and application of an individual’s data.
So what does it mean to “trust” AI in the face of constant, real-world business pressures? And why is trust so important?
What is a trustworthy AI leader?
Our new report, developed with IDC, measures the trustworthiness of AI across five dimensions: Data quality & governance; model governance & oversight; explainability & fairness; responsible AI policy; and audit & accountability. Trustworthiness leaders score high in each of these categories.
Shifting the mindset on AI trust – and the need for safeguards
Closing the AI trust gap is non-negotiable for any organization that expects to gain full value from the technology. Why? Because if you rely on truly trustworthy AI throughout your organization, the AI-powered decisions you make can be trusted as well. But without trustworthy AI principles and the right data foundation, your AI initiatives will fall flat – and the value of AI will erode.
Some feel the framework required to build trusted AI is a constraint on innovation. For example, they may believe it takes too much time to run bias testing or fairness checks and establish governance frameworks. They fear that time spent on those tasks will keep them from getting ahead of competitors in the market.
However, incorporating the necessary AI features to merit trust is what determines whether AI scales and delivers long-term value in high-stakes, real-world situations. Without explainability, governance and ethics, confidence is misplaced, and risks multiply.
When AI systems are transparent and aligned to business goals, leaders and others across their organization will have the confidence to deploy them broadly. Trust then becomes a strategic advantage that accelerates adoption, enables innovation and turns AI into a reliable driver of impact.
Treating trustworthy AI as a repeatable operating model provides a long-term business advantage – and financial returns. It separates those who use AI effectively from those who merely experiment with the technology.
Unfortunately, few organizations today have a centralized team overseeing AI governance or establishing ethics frameworks, fairness checks, data quality monitoring and bias detection to ensure AI is implemented responsibly. Many companies are deploying AI without the safeguards to make it truly trustworthy.
Instead of leaving AI governance to chance – fragmented across teams and systems – it should be a coordinated effort spanning multiple functions. To be effective, governance should include clear accountability, ownership and structured approvals.
Fortunately, we can build systems that help leaders make sound decisions quickly. And we can do it while navigating the moral, operational and financial tensions that create competing priorities, such as reputation, efficiency and cost.
We have to preserve human judgment if we want to preserve human culture. Reggie Townsend Vice President of AI Ethics, Governance and Social Impact SAS
Guiding truths leaders must own to move beyond AI ambition
Leaders who understand the need for a trustworthy AI foundation should embrace three critical tenets to advance their organizations to the next level of AI maturity.
Tenet 1: Trustworthy AI is a leadership responsibility
Like many organizational efforts, trustworthy AI starts at the top. Governance, accountability and ethical guardrails are not just technical concerns. They should reflect leadership’s choices – manifested as mandates for the organization – about how AI is designed, deployed and overseen.
Treating trusted AI as a strategic imperative means establishing clear guidelines and embedding responsible innovation practices across the organization rather than delegating it to individual teams.
Tenet 2: Human oversight matters more as AI becomes more powerful
As AI capabilities become more sophisticated and autonomous, it’s more important than ever for humans to be meaningfully involved. Organizations that treat AI as a replacement for judgment will amplify risk – because AI does not magically eliminate failure.
Trustworthy AI, in the context of human oversight, requires:
- Having transparency, or clarity, into how AI models work.
- Knowing when human judgment is required.
- Understanding how outcomes are monitored over time.
Establishing these guardrails ensures that AI supports human decision-making rather than replacing accountability.
Examples like scaling digital twins in manufacturing with Georgia-Pacific or simulating environments in health care sound AI-first – but they’re not replacing people. They’re bringing people into testing, simulation and decision-making earlier, before anything happens for real.
Tenet 3: Trust protects customer relationships and brand credibility
When AI influences customer interactions, messages and personalization, trust becomes inseparable from brand credibility. It’s crucial for organizations to protect customer data as they would any other valued asset – and to make sure AI-driven customer experiences are fair, explainable and respectful. Because if you lose customer trust, it’s extremely difficult to regain.
Trust is loyalty’s currency in the age of AI. Jenn Chase Executive Vice President and Chief Marketing Officer SAS
Four practices for trustworthy AI leaders
To be successful, AI does not have to be all-knowing – but it does have to be well-governed. Leaders who adopt these four practices can position their organizations to achieve true ROI with their trustworthy AI initiatives.
1. Align AI with clear, strategic goals
The companies that generate tangible ROI aren’t chasing quick wins, they’re thinking bigger. While saving money is often a top goal for AI initiatives, it delivers a lower ROI than other AI goals.
Mature AI organizations (those that have used AI for 8 or more years) prioritize process efficiency and decision-making, with cost savings ranking much lower on their list. Organizations with more strategic AI initiatives significantly expand market share and improve customer experience.
2. Pick the right technology for the job
With emerging AI technologies on the horizon, using AI in isolated ways will leave ROI on the table. The potential of AI – used intentionally, strategically and scaled across the organization – is enormous. Consider that:
- Enterprise agentic AI is expected to transform operations by enabling autonomous decisions, from enhancing customer experiences to flagging fraud.
- Quantum AI promises computational power to tackle complex problems faster.
But we must get comfortable with AI not always being the solution. And don’t forget that traditional AI technologies might often be the best choice. It’s a sign of technological maturity to be able to identify which type of AI is best, and to know when AI is appropriate and when it is not.
To decide which technology to use, start by determining what problem you need to solve – then choose the technology that can best address that issue.
3. Get your data foundation right – be ready for AI
To build trust and support AI innovation, organizations first need a robust data foundation. After all, your data will only work for you if you can trust it, access it and use it effectively.
Clean, centralized and well-labeled data – backed by solid data management practices and supported by AI governance frameworks and ethical oversight – delivers AI outputs that are accurate, auditable and aligned with organizational standards. This is the foundation for establishing AI trust and accountability.
As real-world examples have shown, businesses that rely on insufficient or unreliable training data for large language models (LLMs) risk significant business failures and other bad outcomes.
For example, consider a model built on training data that does not reflect current realities (i.e., it does not account for “data drift” over time). This leads to poor or ineffective business decisions and sometimes creates major damage to an organization’s bottom line and reputation.
4. Prepare for the future you want – and make sure AI serves people
For long-term success with AI, leaders should implement their AI technology end-to-end. As part of this, they should rethink how their data flows, how their teams operate and how their customers are best served – instead of considering only single tools and solutions.
As AI is further embedded in industry workflows, human value will shift toward design, domain expertise, constraint analysis, testing, governance and accountability. All of this is context – and context engineering is becoming the new operating discipline.
AI has the potential to free up human bandwidth to think, design, collaborate and validate at a massive scale, changing outcomes that have been stubbornly stagnant for decades. But as it starts to outperform humans in some ways, it can leave people wondering what value is left for them. Many have started to ask about our future: “Will people matter?”
Always remember that technology should serve people – not the other way around. History shows the only thing that outlasts every innovation is people. So, instead of replacing people with technology, empower people to scale human observation and decision-making.
Empower people with technology to scale human observation and decision-making. Do not replace the human observation; scale it. Bryan Harris Executive Vice President and Chief Technology Officer SAS
Trustworthy AI at SAS
Rather than being a roadblock or just a lofty goal, trustworthy AI empowers us to innovate confidently and ethically.
The Data and AI Ethics Practice at SAS promotes responsible AI use and provides global guidance to prevent harm and build trust. We believe it’s important to ask the hard questions about AI – so, the balance shifts and the questions change.
- “Could we do this?” becomes “Should we do this?”
- “Can we automate this interaction?” becomes “What values are we communicating if we automate that interaction?”
SAS follows a governance model for AI that includes:
- Oversight by a cross-functional executive committee.
- Compliance with evolving global third-party and regulatory frameworks, such as the EU AI Act and those of the National Institute of Standards and Technology.
- Operations that treat trust as a market value – not just a moral one.
- Culture driven by training and education in ethical AI.
AI governance is how we scale our judgment when using AI. It’s not just a technical issue; it’s also a strategic one. Leaders today must consider the moral, operational and financial implications:
- They need AI aligned to what they believe and value most.
- They need the efficiency and productivity of AI without losing quality and consistency
- They need to capture value without creating unmanageable risk.
Those things don’t always line up neatly. But when done well, governance is useful and convenient. It feels like part of the work rather than something layered on top.
Trustworthy AI will reshape our workplaces and society. We must determine how we will intentionally and thoughtfully shape AI's approach to this, in a way that protects what we value. Reggie Townsend Vice President of AI Ethics, Governance and Social Impact SAS
Explore more resources
-
Responsible innovation
Driving technology innovation ethically and responsibly
SAS Trust Center
Security, privacy, compliance and quality for your data
Software
SAS AI Navigator: Enhance AI visibility and reduce shadow AI
