New BARC analyst report reveals what’s missing in enterprise AI trust strategies

Ataccama has released a new report by Business Application Research Center (BARC), The Rising Imperative for Data Observability, which examines how enterprises are building - or struggling to build - trust into modern data systems. Based on a survey of more than 220 data and analytics leaders across North America and Europe, the report finds that while 58% of organizations have implemented or optimized data observability programs – systems that monitor detect, and resolve data quality and pipeline issues in real-time – 42% still say they do not trust the outputs of their AI/ML models.

The findings reflect a critical shift. Adoption is no longer a barrier. Most organizations have tools in place to monitor pipelines and enforce data policies. But trust in AI remains elusive. While 85% of organizations trust their BI dashboards, only 58% say the same for their AI/ML model outputs. The gap is widening as models rely increasingly on unstructured data and inputs that traditional observability tools were never designed to monitor or validate.

Observability is often introduced as a reactive, fragmented, and loosely governed monitoring layer, symptomatic of deeper issues like siloed teams or unclear ownership. 51% of respondents cite skills gaps as a primary barrier to observability maturity, followed by budget constraints and lack of cross-functional alignment. But leading teams are pushing it further, embedding observability into designing, delivering, and maintaining data across domains. These programs don’t just flag anomalies - they resolve them upstream, often through automated data quality checks and remediation workflows that reduce reliance on manual triage. When observability is deeply connected to automated data quality, teams gain more than visibility: they gain confidence that the data powering their models can be trusted.

“Data observability has become a business-critical discipline, but too many organizations are stuck in pilot purgatory,” said Jay Limburn, Chief Product Officer at Ataccama. “They’ve invested in tools, but they haven’t operationalized trust. That means embedding observability into the full data lifecycle, from ingestion and pipeline execution to AI-driven consumption, so issues can surface and be resolved before they reach production. We’ve seen this firsthand with customers – a global manufacturer used data observability to catch and eliminate false sensor alerts, unnecessarily shutting down production lines. That kind of upstream resolution is where trust becomes real.”

The report also underscores how unstructured data is reshaping observability strategies. As adoption of GenAI and retrieval-augmented generation (RAG) grows, enterprises are working with inputs like PDFs, images, and long-form documents – objects that power business-critical use cases but often fall outside the scope of traditional quality and validation checks. Fewer than a third of organizations are feeding unstructured data into AI models today, and only a small fraction of those apply structured observability or automated quality checks to these inputs. These sources introduce new forms of risk, especially when teams lack automated methods to classify, monitor, and assess them in real time.

“Trustworthy data is becoming a competitive differentiator, and more organizations are using observability to build and sustain it,” said Kevin Petrie, Vice President at BARC. “We’re seeing a shift: leading enterprises aren’t just monitoring data; they're addressing the full lifecycle of AI/ML inputs. That means automating quality checks, embedding governance controls into data pipelines, and adapting their processes to observe dynamic unstructured objects. This report shows that observability is evolving from a niche practice into a mainstream requirement for Responsible AI.”

The most mature programs are closing that gap by integrating observability directly into their data engineering and governance frameworks. In these environments, observability is not siloed; it works in concert with DataOps automation, MDM systems, and data catalogs to apply automated data quality checks at every stage, resulting in improved data reliability, faster decision-making, and reduced operational risk.

Ataccama partnered with BARC on the report to help data leaders understand how to extend observability beyond infrastructure metrics or surface-level monitoring. Through its unified data trust platform, Ataccama ONE, organizations can apply anomaly detection, lineage tracking, and automated remediation across structured and unstructured data. Observability becomes part of a broader data trust architecture that supports governance, scales with AI workloads, and reduces the operational burden on data teams.