DATA AND AI IMPACT REPORT: The Trust Imperative

In just the last two years, generative AI has already eclipsed the use of traditional AI. As the market rapidly advances to agentic AI, the impact on decision making will be pervasive, often concealed behind automation and integration.

For the good of society, businesses and employees: Trust in AI is imperative.

In order to achieve trust, the AI industry must increase the success rate of implementations, humans must critically review AI results and leaders must empower the workforce with AI.

We set out to tackle these important topics and other questions in a global survey around the use, impact and trustworthiness of data and AI.

In these pages, you’ll learn:

We tend to overly trust more humanlike technology.

In spite of evidence that GenAI can be error-prone, organizations have more faith in this technology than in other types of AI, including machine learning.

Tangible ROI increases with trust.

To realize greater value from AI initiatives, the survey found businesses must focus on governance, explainability and ethical safeguards.

Businesses must move beyond cost cutting. 

Saving money is often a top goal when it comes to AI initiatives, but the study showed it delivered the lowest ROI compared to other AI goals. Organizations with more strategic AI initiatives significantly expand market share and improve customer experiences.

Agentic AI requires groundwork.

The study revealed that agentic AI progress will likely stall when faced with non-optimized cloud data environments, poor data governance, or talent shortages.

Quantum AI is quickly becoming a reality.

Quantum applications are currently being explored in logistics, finance, cybersecurity, life sciences, climate modeling, and materials science. Although this technology is in an experimental phase, survey respondents shared that the excitement for its potential is very real.

In addition to these findings, you will also learn about the AI Trust Index and the AI Impact Index – two new measurements we are introducing to compare the use and impact of AI across regions and industries.

For AI to reach its next growth phase, it must be grounded in delivering tangible ROI to businesses.

This requires everyone to overcome the trust dilemma.

Dave Schubmehl
Research Vice President,
AI & Automation, IDC

What stands out in this research is how rapidly the center of gravity has shifted from traditional machine learning toward generative and agentic AI. Organizations are no longer experimenting on the edges; they are embedding these technologies into workflows that span customer service, coding and decision support. Yet, the real differentiator is not just adoption but integration: the ability to unify structured and unstructured information, apply governance and embed explainability into automated processes. Companies that align trustworthy AI practices with their automation strategies are seeing the clearest improvements in efficiency and workflow impact, while those that don’t risk amplifying inefficiencies at scale.

Kathy Lange
Research Director,
AI Software, IDC

What this research makes clear is that the conversation about AI trust and impact cannot be separated from the AI life cycle itself. From data preparation and labeling to model training, deployment and monitoring, every stage introduces opportunities and risks that directly affect outcomes. Too often, organizations focus on adoption metrics without investing in life cycle automation, governance and continuous oversight – leaving them exposed to bias, drift or compliance failures. The data shows that those building maturity across the full AI pipeline not only solve the trust dilemma, but also realize greater business value. Trustworthy AI is not a final checkpoint; it is embedded in every phase of the life cycle.

To understand how organizations approach and benefit from AI, this report introduces three core metrics. The Trustworthy AI Index captures the strength of practices that ensure AI systems are built responsibly and deserving of trust. The Impact Index measures the tangible business value organizations derive from AI investments. Finally, the concept of the trust dilemma highlights where trust and trustworthiness fall out of alignment, exposing either underutilized potential or elevated risk. Together, these measures provide a comprehensive view of how AI is governed, how it delivers impact and where organizations must act to align trust with outcomes.

The chart below plots the relationship between the perceived trust in AI systems and their actual trustworthiness, illustrating the “trust dilemma.” This misalignment, evident across all regions, represents a critical barrier to effective AI adoption. Most organizations experience this misalignment, with relatively few achieving the ideal balance. Two risks emerge: underutilization of reliable systems when trust remains low and overreliance on unproven systems when confidence is disproportionately high. The challenge is particularly acute for generative AI, where rapid enthusiasm has outpaced governance and data quality.

The trust dilemma is a persistent global issue, affecting nearly half of organizations worldwide (46%). It is slightly more pronounced in Asia Pacific and North America, where 47% of organizations face misalignment between trust in AI and actual system trustworthiness. META (Middle East, Turkey and Africa) and Latin America perform slightly better at 45%, though the dissonance remains significant. Even in Europe, where regulatory oversight is stronger, 46% of organizations still fall into this dilemma. Solving it will require sustained investment in governance frameworks, the development of skilled talent and robust infrastructure to ensure that organizational confidence in AI technologies is firmly grounded in demonstrable reliability and integrity. Ultimately, solving this trust dilemma is essential if organizations are to unlock the full impact and value of AI.

Organizations that get the most ROI from AI have more strategic goals

Companies that get the most ROI from AI aren’t just trying to cut costs—they’re reimagining how their organizations work.

This research shows that newer AI adopters (less than two years in) tend to focus on personal productivity, with 57% listing it as their top priority. But in organizations with over eight years of AI experience, the focus shifts to process efficiency (64%) and decision making (60%). These mature users are the only ones who consistently prioritize decision making as a key AI goal.

The data also reveals that cost reduction, while common among newer adopters, delivers the lowest ROI. In fact, saving costs ranks as the third-highest priority for AI newcomers, but drops to seventh for mature organizations. Companies that aim to improve customer experience, expand market share and boost business resilience report significantly higher returns. This shows that focusing on using AI to enhance processes and deliver better experiences across the board is key to unlocking measurable value.

From GenAI to agentic AI: The next leap requires a stronger foundation

Agentic AI is poised to revolutionize enterprise operations, but most organizations aren’t ready for it. Like GenAI, agentic AI is being held back by weak data infrastructure, poor governance and a lack of AI skills. 

To scale value from these technologies, companies must build strong data foundations that support innovation. Right now, 49% of organizations cite nonoptimized cloud data environments as a major barrier, followed by insufficient data governance (44%) and a shortage of skilled specialists (41%). For more experienced AI organizations, the challenges shift toward building organizational AI architecture (57%), running effective data and AI teams (45%) and supporting training and reskilling (42%).

Since 2024, data centralization has surged to become the number one challenge in AI implementation. Concerns around data privacy, explainability and ethical use are also top of mind for organizations deploying agentic AI.

With these limits in mind, agentic AI adoption will be slower than GenAI because it requires organization-wide process redesign and scales one agent at a time. As companies progress in AI maturity, they’ll face new hurdles like API integration, agent orchestration and hybrid cloud data management. Executives focused on process efficiency must recognize that a solid data foundation is essential for driving impact with agentic AI.

Quantum AI Is gaining confidence faster than capabilities

Quantum AI is gaining trust faster than it is gaining understanding. According to our findings, 30% of global decision makers say they’re familiar with quantum AI, and 25% say they trust it, despite real-world applications still being rare and mostly confined to research labs. Global leadership in this space is concentrated in the US, Europe and select Asian innovation hubs, but the technology remains in its early stages, defined by experimental hardware and a handful of promising research-led use cases.

This research reveals that while quantum AI is on the threshold of real-world impact, its adoption is largely aspirational. The fusion of quantum computing and AI holds transformative potential in fields like logistics, finance, cybersecurity and climate modeling. To prepare, organizations should explore hybrid quantum-classical architectures, where quantum processors work alongside traditional AI systems to accelerate complex model training and inference. As advancements in quantum hardware and error correction continue, strategic attention is essential – even if full-scale deployment is still a future goal.

Despite being the newest and trendiest form of AI, generative AI is the most trusted – by a wide margin. Survey respondents scoring low on trustworthiness trust GenAI 200% (3x) more than traditional machine learning, even though machine learning is mathematically explainable and far more transparent (blue outlined boxes). This paradox reveals a humanlike bias: People tend to trust AI that feels intuitive and conversational, even when they know it may not always be accurate. The more “human” an AI feels, the more we trust it, regardless of its actual reliability.

At the same time, trust in GenAI is layered with concern. While users report high levels of trust, they also express significant worries about data privacy (62%), transparency and explainability (57%) and ethical use (56%). This trust is often driven by perceived usefulness and interactivity – GenAI delivers fast, tailored responses that feel helpful, especially in low-stakes situations. But interactivity doesn’t equal control, and as GenAI becomes more embedded in daily workflows, organizations must be cautious not to mistake familiarity for reliability.

As artificial intelligence continues to reshape industries, the level of trust in AI and the investment in trustworthy AI practices vary significantly across sectors. From banking and insurance to government and life sciences, organizations are navigating the complex balance between innovation and responsible adoption.

Solving the trust dilemma

While each sector faces unique challenges, a common theme emerges: Trust in AI must be matched by tangible investments in governance, talent and infrastructure. Solving the trust dilemma is not just a matter of belief in AI’s potential – it requires deliberate action to ensure responsible and sustainable adoption. To solve the trust dilemma, organizations must focus on strengthening data and model governance, embedding ethical and transparent practices into AI life cycles and developing the skills and culture needed to align human trust with technological trustworthiness.

Among the four industries examined in this study, life sciences stands out for its advanced maturity in both AI and data infrastructure. A significantly higher share of life sciences organizations identify as being in the “transformative” stage of AI maturity compared to banking, insurance and government. This leadership position also extends to data infrastructure maturity.

OVERALL TRUSTWORTHY AI:
TRANSFORMATIONAL STATUS

TRUSTWORTHY AI INVESTMENT FOR
AGENTIC AI: SIGNIFICANTLY INCREASING

Life sciences organizations outperform only the government sector in their current efforts to deliver trustworthy AI, with 20% operating at the highest level of our Trustworthy AI Index – slightly above the global average of 19.8%. However, they are the least likely among the four industries to anticipate a significant increase in future investment toward trustworthy AI initiatives that support agentic AI, falling behind both our other focus industries and the global average.

Life sciences organizations generally have a long-standing history of working with AI. Given the nature of the industry, they tend to approach AI implementation with a strong emphasis on responsibility and ethical use to ensure safety and meaningful outcomes for all patients.

Globally, life sciences organizations report facing key AI implementation challenges, particularly around data foundations, a lack of specialized AI technology personnel and data governance. In response, their top priorities include building appropriate data and technology architectures and investing in the development of specialized teams to address these gaps.

Among our four focus industries, life sciences organizations express the most modest expectations for future AI investment. Under 8% expect investments in AI to grow greater than 20% over the coming year, and over a third (34.4%) expect only a limited change in AI investment in that period.

When it comes to AI’s business value, nearly two-thirds of life sciences leaders prioritize process efficiency and effectiveness as the primary outcome of AI adoption.

Among the four industries analyzed in this study, insurance presents the most modest profile in terms of both AI and data infrastructure maturity. Insurance organizations report the lowest share of respondents identifying as being in the “transformative” stage of AI maturity compared to government, life sciences and banking. This trend is mirrored in data infrastructure maturity as well.

Insurance organizations outperform life sciences and government sectors in their current efforts to deliver trustworthy AI, with 20.1% operating at the highest level of our Trustworthy AI Index – slightly above the global average of 19.8%. Only banking ranks higher in this regard. When it comes to future investment plans for trustworthy AI initiatives supporting agentic AI, insurance organizations are roughly aligned with the global average, again trailing only banking respondents.

Globally, insurers express modest expectations for AI investment growth. Fewer than 8% anticipate a significant increase in investment over the next year, while more than 30% expect only limited changes in investment. Nearly two-thirds of insurance leaders identify process efficiency and effectiveness as the primary way AI delivers business value.

Like the broader respondent base, insurers cite data foundations, data governance and limited access to specialized AI technology personnel as major implementation challenges. Notably, insurers are the most likely among the four industries to identify data governance as a top concern.

In response to these challenges, insurers are prioritizing skill development across both general employee populations and technical specialist roles focused on data and AI. Compared to other industries, insurers are slightly less likely to emphasize architectural solutions as key implementation priorities to address data governance issues going forward. Additionally, about one-third of insurance organizations are prioritizing the creation of AI strategies and roadmaps.

Banking organizations show modest levels of maturity in both AI and data infrastructure. The proportion of banking respondents identifying as being in the “transformative” stage of AI maturity is roughly on par with insurance and government organizations, with only life sciences reporting higher levels.

Banking organizations lead all other focus industries in their current efforts to deliver trustworthy AI, with 23.4% operating at the highest level of our Trustworthy AI Index – well above the global average of 19.8%. They are also significantly more likely than peers in life sciences, insurance and government to anticipate increased future investment in trustworthy AI initiatives that support agentic AI.

Globally, banking organizations report the most aggressive investment outlook for AI among the four industries – 11.5% expect AI investment to grow by more than 20%, and nearly 60% anticipate growth between 4% and 20%. Unlike other sectors, banking leaders prioritize product and service innovation over process efficiency and effectiveness as the primary source of AI-driven business value.

Banking organizations identify key challenges in AI implementation, including a lack of data governance, a lack of specialized AI technology personnel and insufficient data centralization and optimization. However, it’s notable that none of these issues were cited by half or more of banking respondents.

In response to these challenges, banking organizations are prioritizing the development of robust technology architectures and workforce skills. Notably, they also rank the development and tuning of AI models among their top five implementation priorities.

Government organizations report notably strong AI maturity on a global scale. Nearly half of respondents place themselves in either the “transformative” or “integrated” stages of maturity – significantly ahead of banking and insurance organizations. However, the picture shifts when it comes to data infrastructure maturity: Government organizations trail life sciences by a wide margin, though their responses are generally consistent with those from banking and insurance.

In terms of delivering trustworthy AI, government organizations lag behind the other three focus industries. Only 15.3% operate at the highest level of our Trustworthy AI Index, compared to the global average of 19.8%. They also fall short of banking and insurance organizations in their expectations for future investment in trustworthy AI initiatives.

Like the other industries in this study, government organizations frequently cite challenges related to data foundations, data governance and skills availability. However, they stand out as the only sector where respondents are more likely to highlight skills gaps among general employee populations rather than among specialized technical teams.

Reflecting these challenges, government organizations are prioritizing investments in technology architecture alongside workforce skill development. Nearly one-third of respondents also emphasize the importance of creating an AI strategy and roadmap.

Government organizations express strong expectations for AI investment growth in the coming year – 12.6% anticipate increases of more than 20%, and nearly half expect growth between 4% and 20%. Like leaders in life sciences and insurance, government respondents view process efficiency and effectiveness as the primary lens for realizing AI’s business value. Additionally, personal productivity is cited by over 60%, the highest rate among all four industries.