Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Andrew Crisp, SVP and Global Data Strategy Leader, Dun & Bradstreet
As AI becomes embedded into enterprise workflows, organizations are discovering something fundamental: Large models will confidently amplify whatever they are fed. Without trustworthy data, AI will produce mistakes at scale. The more enterprises attempt to automate, the more visible this truth becomes.
In 2026, AI maturity will become inseparable from data maturity, and verification will move from a back-office concept to a front-and-center priority. As a result, enterprise adoption of AI will hinge less on breakthroughs in AI models and more on whether the underlying data can be verified and trusted.
AI Will Not Fix Bad Data
Responsible enterprise adoption of AI starts with the acknowledgement that AI magnifies bad data, not fixes it. If identity resolution is weak, agents become unreliable. If sources are unverified, models become unpredictable. Business leaders are seeing this firsthand as they push AI deeper into revenue, finance, compliance, procurement, and operations.
For many executives, hallucinations were once a curiosity. Entertaining, even. That ended when AI systems began handling nuanced decisions and communications that required human oversight. Once models became intertwined with regulatory exposure and financial risk, the tolerance for inconsistency dropped to zero.
In 2026, the most strategic question will shift from “Which model should we use?” to “How do we ensure every model sees trustworthy, verified data?”
What It Takes to Power AI Reliably
As we move forward, enterprises are consolidating around a clearer definition of what reliable AI requires. These requirements include:
1. A verification layer grounded in identity. AI agents need to know precisely which entity they are acting on. Ambiguity breaks automation. A trusted identity graph, complete with hierarchies and linkages, closes this gap.
2. A living entity graph, not a static database. AI systems operate continuously. Their underlying data must do the same. Organizations are beginning to shift from periodic data refreshes to continuously updated, explainable entity graphs that ingest structured and unstructured signals and reflect real-world change in near-real time.
3. Auditable and compliant inputs. Enterprises cannot scale AI without robust controls. Every attribute must map back to a governed standard. Without this, risk teams slow or blocked deployments.
4. Context-rich, structured delivery. Models perform best when they are fed structured, contextualized data rather than raw inputs. As agentic AI implementations proliferate, standardized formats, including emerging protocols for model context and agent-to-agent communication, will become essential.
5. The fusion of LLM reasoning with authoritative reference data. Model reasoning alone does not produce enterprise-grade precision. LLMs are powerful, but they must be paired with authoritative reference data.
From Data Pipelines to Data Verification Pipelines
As organizations move deeper into AI, they are discovering that traditional data pipelines, built largely for reporting, are insufficient. AI agents do not consume data the way dashboards do. They require structured identity and persistent context.
This is why 2026 will see a significant rise in data verification pipelines designed specifically for AI consumption. Instead of focusing on extract-transform-load mechanics, enterprises will increasingly emphasize:
- Identity resolution at speed
- Automated anomaly detection
- AI-assisted enrichment of missing attributes
- Verification of external signals
- Continuous quality monitoring
- Real-time governance and lineage capture
Preparing for This Shift
As organizations move toward agentic systems and automated decisioning, they will need a backbone of verified data. With this in mind, Dun & Bradstreet is investing in AI tools that strengthen the quality and timeliness of the data itself. We are applying AI to ensure that our global data becomes a living entity graph: a continuously updated, verifiable identity fabric that can confidently power enterprise AI at scale.
This is the one place where our own philosophy is worth stating directly: Trusted data is not static. It must evolve with the pace of the real world. AI allows us to do that with greater speed and precision than ever before.
Trusted and Verified Data as the Differentiator
In 2026, AI success will depend less on choosing the right model and more on whether the system is held up by a strong data backbone. Organizations that treat data verification as the new foundation of AI will find the fastest time to value. Those that skip this step will continue to struggle with hallucinations and stalled deployments.
The promise of AI is real, but it is only realized when the underlying data can be trusted. Verification is now the central requirement for AI that enterprises can rely on, and it will be the defining characteristic of the next phase of AI maturity.
##
ABOUT THE AUTHOR
Andy Crisp is Dun & Bradstreet�s Senior Vice President, Global Data Owner, leading the Global Data Owners, Global Data & Trade Acquisition, and Data Quality Insights functions. With nearly 25 years at D&B, Andy is responsible for the stewardship, sourcing, and strategic management of the data that powers the company�s commercial intelligence across more than 220 countries. He oversees global data strategy execution and drives initiatives that ensure the quality and accessibility of D&B�s critical data assets.
A seasoned data leader, Andy has built his career across a range of disciplines including sales, project management, data management, and data strategy. Over the years, he has held progressively senior roles across core data functions, with a consistent focus on building high-performing global teams, leading complex transformation programs, and advancing the value of data across the organization.
Andy is widely recognized as a thought leader in data quality and innovation, particularly in the practical application of artificial intelligence (AI) and machine learning (ML) technologies into D&B�s data operations�enhancing data quality,coverage, and decisioning across global markets. His leadership has enabled D&B to deliver trusted, scalable data solutions to clients worldwide.
Andy is a four-time honouree of DataIQ�s UK Top 100 most influential people in data, reflecting his sustained impact on the industry and commitment to advancing data-driven business.
From 2020 to 2025, Andy served as a Committee Member of the Data, Analytics and AI Leadership Committee at TechUK, where he played a key role in shaping strategic priorities to accelerate data-driven innovation and AI adoption across the UK�s public and private sectors.
