Trust gaps with AI may be hindering business growth

The Trust Gap in AI: Why Businesses Are Struggling to Realize Its Potential

Artificial intelligence (AI) has become a central focus for businesses around the world. Many leaders recognize that AI is essential for future growth and competitiveness. However, despite the excitement surrounding AI, many organizations still struggle to trust their AI systems or see tangible results from them. Two recent studies highlight the challenges companies face in fully leveraging AI and offer insights into how they can overcome these obstacles.

Distrust of AI Starts With Data Quality

A report by data management company Ataccama revealed that 42% of organizations do not trust the outputs of their AI or machine learning models. This is concerning given the significant investments being made in AI initiatives. The root cause, according to the study, often lies in poor data quality. Only 58% of businesses have implemented or optimized data observability programs—processes that allow companies to monitor, check, and improve data quality across all stages.

Data observability is more than just looking at dashboards. It involves tracking data throughout its entire lifecycle, from initial ingestion to processing and consumption by AI models. Jay Limburn, Chief Product Officer at Ataccama, emphasized that many companies invest in tools but fail to operationalize trust. This means embedding observability into every stage of the data lifecycle so issues can be identified and resolved before they impact production.

Many businesses take a piecemeal or reactive approach, addressing problems only after they lead to faulty outputs or system failures. This fragmented strategy creates data silos and weak governance, which increase the likelihood of unreliable AI results.

The Challenge of Unstructured Data

Another growing issue is the handling of unstructured data, such as PDFs, images, audio files, and free-text documents. Traditional data observability tools are often not designed to manage this type of data. However, with the rise of generative AI and retrieval-augmented generation (RAG) approaches, more businesses are working with unstructured data than ever before.

Ataccama noted that less than one-third of organizations currently use unstructured data in their AI models. This represents both a missed opportunity and a risk, as even when companies do incorporate unstructured data, their systems may lack the capability to monitor its quality and consistency effectively.

Advanced companies are addressing this gap by integrating observability directly into their data engineering and governance frameworks. They also implement automated data quality checks and remediation workflows to ensure that any data issues are flagged and resolved before they affect AI models.

When done correctly, this approach leads to more reliable data, faster decision-making, and reduced operational risks.

Skills, Budgets, and Organizational Trust Are Major Barriers

A separate global survey by Qlik, which included over 4,000 C-suite executives and AI decision-makers, uncovered additional challenges hindering AI adoption. While 88% of business leaders believe AI is critical to achieving their goals, many AI projects remain stuck in planning stages.

For example:
– 20% of businesses have between 50 to 100+ AI projects in the planning phase that haven’t launched.
– Another 20% have up to 50 projects that were paused or canceled during planning.

The main reasons for this stagnation include:
– Lack of AI skills: 23% struggle with developing AI, and 22% lack people who can deploy it properly.
– Data governance issues: 23% cited this as a top roadblock.
– Budget constraints: 21% say funding is a limiting factor.
– Untrusted data: Also cited by 21%.

Even more troubling, over a third (37%) of AI decision-makers reported that senior managers don’t trust AI, and 42% felt that less senior employees also lack confidence in AI. Additionally, 21% believe customers are wary of AI as well. As a result, 61% of companies said trust issues are causing them to scale back AI investments.

How Companies Can Build Trust and Move AI Forward

Experts suggest that building trust starts with transparency, training, and focusing on clear, achievable use cases. James Fisher, Qlik’s Chief Strategy Officer, emphasized the importance of defining a specific goal, such as improving customer churn prediction or streamlining supply chains, and then mapping out the required data, skills, and resources. This helps build confidence and gain support from leadership.

Qlik’s research also found that:
– 74% of AI leaders want to promote the benefits of AI within their organizations and to customers.
– 76% believe their industries need better staff upskilling for AI.
– 75% want governments to provide more funding and training programs in AI.

Many companies are also turning to ready-made AI solutions that can be tailored to their needs instead of building everything from scratch. This reduces technical barriers and accelerates time to value, making it easier to demonstrate ROI and justify further investment.

Key Takeaways for Building AI Trust and Success

To unlock the true potential of AI, businesses must address several key areas:

1. Prioritize Data Quality

Strong, proactive data observability across the entire lifecycle is crucial. Without it, even the best AI tools will fail to earn trust.

2. Handle Unstructured Data Effectively

As unstructured data grows, companies need systems capable of monitoring and ensuring the quality of text, images, and other non-tabular data.

3. Invest in Skills and Governance

Upskilling employees and strengthening data governance is essential to prevent AI projects from stalling or being abandoned.

4. Foster Organizational Trust

Trust is a hidden but vital component of AI success. Transparent communication and small, measurable wins can help build confidence and justify continued investment.

5. Start with Focused Use Cases

Pick projects with clear goals and success metrics. Begin with what’s achievable and build from there.

6. Consider Ready-Made Tools

Off-the-shelf AI solutions can help bypass planning paralysis and deliver quick returns, proving the value of AI investment.

By addressing these underlying issues—especially data quality and organizational trust—businesses can finally unlock the full power of AI and turn ambitious promises into real, measurable outcomes.