Building Trust Before Intelligence: How Sunil Kanojiya's Governance-First Framework Is Helping Organizations Create Trustworthy AI

From hospital systems to wholesale distribution, organizations across every sector are collecting more data than ever before, and trusting it less. A new generation of data governance specialists is changing that equation, one pipeline, one framework, and one dashboard at a time.

There is a problem hiding inside most organizations, and it does not appear on any financial statement. It does not trigger an alarm or generate a ticket. In many cases, the people closest to it do not even know it exists. The problem is this: the data those organizations rely on to make decisions, every day, at every level, cannot actually be trusted.

This is not a storage problem or a software problem. It is a governance problem. Data is being collected at unprecedented scale across healthcare systems, enterprise software platforms, supply chains, and wholesale operations. Yet in organization after organization, the data flowing through dashboards, financial models, and clinical systems is inconsistent, duplicated, poorly defined, and stripped of the context that would make it genuinely useful. The result is a gap between what organizations think their data tells them and what it actually means.

Sunil Kanojiya, a Data and Analytics Manager with experience spanning Workday Inc., HCA Healthcare, UnitedHealth Group, and ABS Wholesale, has spent his career working inside that gap. His approach, combining technical data engineering with rigorous governance frameworks and published academic research, offers a template for how organizations can close it.

The Scope of the Problem Is Larger Than Most Executives Realize

Most senior leaders understand that data quality is important. Fewer understand how comprehensively poor governance undermines the systems they depend on. When a financial model produces a forecast, when a clinical dashboard flags a patient risk, when a supply chain system recommends a procurement decision, the quality of that output is only as strong as the governance of the data that produced it. Garbage in, garbage out is one of the oldest expressions in computing. It remains one of the most consistently ignored.

Kanojiya encountered this pattern repeatedly across his career. At Workday Inc., working inside one of the world's most sophisticated enterprise software environments, the technical infrastructure was state of the art. But without governance frameworks that enforced consistent data definitions, standardized ETL workflows, and documented data ownership, the dashboards those systems produced still told different stories to different teams. Two departments could pull reports from the same system on the same day and arrive at entirely different conclusions about the same metric.

The pattern repeated at HCA Healthcare, one of the largest hospital systems in the United States, operating 186 facilities across 20 states. Here, the stakes were not financial performance or operational efficiency alone, they were clinical. Patient data flowing across laboratory systems, electronic health records, billing platforms, and imaging systems carried inconsistencies that could, if unaddressed, affect diagnosis and treatment. The work Kanojiya contributed to, developing Power BI and Tableau dashboards that unified clinical and operational KPIs, and deploying predictive analytics to identify cost drivers in care delivery, contributed to a documented 20 percent improvement in claims processing efficiency, a figure Kanojiya attributes directly to the governance layer, not the analytics tools themselves. The underlying insight was consistent across his tenure there: the technology was capable. The governance was the constraint.

What Data Governance Actually Means in Practice

The term data governance has accumulated a great deal of baggage. In many organizations, it is associated with compliance checklists, IT projects that run over budget, and frameworks that get documented and then ignored. This association has made many business leaders skeptical of governance as a practical investment. Kanojiya's central argument, developed through years of applied work and formalized in his book Engineering Trust in Data, is that this association is the product of how governance has been implemented, not what governance actually is.

Real governance is not a compliance exercise. It is an operational capability. It defines which data elements mean what, who owns them, how they are validated before they enter downstream systems, how inconsistencies are detected and resolved, and how the chain of custody, from source to dashboard, is made visible and auditable. When governance is built into a data environment from the outset, it becomes invisible in the best possible sense. Analysts spend their time generating insights rather than reconciling conflicting numbers. Executives trust the dashboards in front of them. Clinical staff rely on the information their systems surface. Supply chain managers act on inventory signals rather than discounting them.

At ABS Wholesale, Kanojiya built this infrastructure from scratch. The company had no formal governance structure when he arrived. What it had was data, collected from sales, procurement, inventory, and logistics systems, and no reliable way to make sense of it. Over the following months, he designed and implemented a comprehensive data governance framework: data dictionaries that standardized definitions across departments, validation rules embedded in daily ETL workflows, automated anomaly detection that flagged inconsistencies before they reached reporting layers, and KPI dashboards that gave leadership a unified, reliable view of operational performance. The result was not just cleaner data. It was a fundamentally different relationship between the organization and the information it generated.

Predictive Modeling: From Historical Reporting to Forward-Looking Intelligence

The progression from descriptive analytics, what happened, to predictive analytics, what is likely to happen next, is one of the most significant shifts available to data-mature organizations. It is also one of the most commonly misunderstood. Many organizations invest in predictive tools before their data is ready to support them, and then wonder why the models fail to deliver. The reason is almost always the same: predictive modeling built on poorly governed data produces authoritative-looking outputs that cannot be trusted.

This is a point Kanojiya has made explicitly in his published research. His paper Reducing Healthcare Costs Using Predictive Modeling: A Data-Driven Framework for Optimizing Clinical and Operational Efficiency, published in The American Journal of Applied Sciences in May 2026, proposes a structured framework for deploying predictive models in healthcare settings. The framework is notable for what it leads with: not the modeling techniques themselves, but the governance preconditions that must be in place before those techniques can produce reliable results. Clean inputs. Validated schemas. Documented lineage. Defined ownership. These are not preliminary steps that can be skipped. They are the foundation on which every downstream insight depends.

His complementary paper, Regulatory-Compliant Data Analytics: Designing HIPAA-Aligned Data Pipelines at Scale for Secure and Efficient Healthcare Data Processing, addresses the compliance dimension of the same challenge. In healthcare, data governance is not optional, it is mandated. HIPAA requires organizations to demonstrate how protected health information is collected, processed, transmitted, and stored. Building analytics pipelines that meet these requirements without sacrificing the performance and flexibility that clinical and operational teams need is a genuine engineering challenge. The paper lays out an architecture for doing so at enterprise scale.

What makes Kanojiya's research unusual is its grounding in real deployments rather than theoretical models. The frameworks he proposes have been tested in the environments he describes, healthcare systems and enterprise operations where the cost of a mislabeled field is not an abstraction but a claims denial, a budget shortfall, or a clinical error.

The Cross-Industry Pattern: Healthcare, Enterprise, and Distribution

One of the most instructive aspects of Kanojiya's career is its breadth. Most data practitioners develop deep expertise in a single domain. Kanojiya's trajectory spans healthcare, enterprise software, and wholesale distribution, three environments that appear superficially different but share a remarkably consistent underlying data challenge.

In healthcare, data fragmentation is driven by the proliferation of specialized systems, EHRs, laboratory platforms, imaging systems, billing platforms, that were built independently, often by different vendors, and have never been fully integrated. The clinical consequences of this fragmentation are direct: incomplete patient histories, inconsistent clinical coding, delayed diagnostic data. The operational consequences are equally significant: billing errors, claims denials, compliance exposure. Governance in this environment requires not just technical integration but semantic alignment, ensuring that a patient identifier, a diagnostic code, or a procedure record means the same thing in every system it appears in.

In enterprise software environments like Workday, the fragmentation takes a different form. The systems are more integrated by design, but the organizational structures that use them are not. Finance, HR, operations, and sales teams often develop their own interpretations of shared data, leading to the kind of metric inconsistency that makes cross-functional reporting unreliable. Governance here is primarily a definitional and cultural challenge: getting departments to agree on what their shared data means and maintaining that agreement as systems and strategies evolve.

In wholesale distribution, the challenge is operational velocity. Inventory data, procurement records, logistics information, and sales figures move quickly and interact with one another in ways that require real-time consistency to be useful. An inventory signal that is twelve hours out of date when a procurement decision is being made is not just imprecise, it is actively misleading. Governance in this context means embedding quality checks and validation into the pipelines themselves, not reviewing data after it has already influenced decisions.

What connects these three environments is the same fundamental constraint: organizations have invested in data collection and in analytics tools, but they have underinvested in the governance layer that makes those investments coherent. Kanojiya's work, across all three domains, has been to build that layer.

The AI Readiness Problem No One Is Talking About

The conversation about artificial intelligence in enterprise settings has accelerated dramatically over the past two years. Executives are being told, from every direction, that AI will transform their operations, automate their workflows, and generate insights that no human analyst could produce. What those conversations rarely address is the precondition that determines whether any of this is true for a given organization: the quality and governance of the data those AI systems will be trained on and will operate against.

AI does not fix bad data. It amplifies it. A machine learning model trained on inconsistent, poorly governed data will learn the inconsistencies as well as the patterns. A predictive model built on data that has not been validated for completeness, accuracy, and lineage will produce predictions that are statistically sophisticated and practically unreliable. The organizations currently racing to deploy AI capabilities without first addressing their governance foundations are building on unstable ground, and many of them do not yet know it.

This is a point Kanojiya has made in his research and in his work at ABS Wholesale, where he has begun integrating AI-assisted analytics into operational workflows. The sequence matters: governance first, then AI. Clean, well-defined, consistently validated data is not just better input for AI systems. It is the condition that makes the outputs of those systems trustworthy enough to act on. Organizations that get this sequence right will realize genuine value from their AI investments. Those that skip the governance step will generate impressive-looking models that their own teams will eventually stop trusting.

The Peer Review Dimension: Expertise Recognized Beyond the Employer

One of the less visible but significant aspects of Kanojiya's profile is the recognition his expertise has received from academic and research organizations outside his employment context. The Institute for Technology and Research, the International Journal of Industrial Electronics and Electrical Engineering, and Science Guru have all invited him to serve on their editorial committees and as a peer reviewer for conference and journal submissions. In early 2026, he completed formal peer reviews for manuscripts on AI-powered information management and IoT-based security systems, submitting detailed assessment reports to the conference organizing committee.

This form of recognition matters because it did not originate with an employer, a client, or Kanojiya himself. It was extended by independent research bodies with no stake in his career. An employer letter describes what a professional does within the organization. A peer review invitation describes what the broader research community believes that professional is qualified to evaluate. When multiple international research organizations independently conclude that a practitioner's expertise is authoritative enough to assess the work of other scholars, it reflects a level of field recognition that extends well beyond job performance.

His Project Management Professional certification from the Project Management Institute, one of the most demanding credentials in the technology sector, held by fewer than one million practitioners globally, adds another dimension to this picture. The PMP is not a credential that can be earned by accumulating experience alone. It requires documented evidence of project leadership, a rigorous examination, and a formal commitment to ongoing professional development. It signals not just competence but a level of systematic, organized thinking about how complex initiatives are planned and executed.

Engineering Trust: The Book and the Argument

Engineering Trust in Data, the book Kanojiya authored alongside co-researchers Sadia Afroz, Hasib Ur Rashid, MD Al-Amin Chowdhury, and Shuvo Ranjan Das, is the most complete articulation of his central argument. The book runs to twelve chapters covering the full lifecycle of data governance: from the trust crisis that motivates the need for governance, through the technical architecture of trusted data systems, to the organizational transformation required to sustain governance culture over time. It is not a theoretical text. It is a practitioner's framework, informed by deployment experience across multiple industries and validated by the research literature on data management, governance, and analytics.

The book's central claim, that data quality is not primarily a technical problem but a governance problem, and governance is ultimately a leadership problem, is both simple and consequential. Organizations that treat data quality as an IT matter will continue to find that their IT investments do not solve it. Organizations that treat it as a governance matter, requiring executive ownership, clear role definition, standardized processes, and embedded cultural norms, will find that relatively modest investments in the right places produce dramatic improvements in the reliability and utility of their data.

This is the argument that underpins all of Kanojiya's work, across all the organizations he has served and all the research he has published. The data exists. The technology exists. What most organizations lack is the governance layer that makes both of those things trustworthy. Building that layer is the work. And it is work that has measurable, demonstrable consequences for every industry that depends on data to function, which, in 2026, is every industry.

Looking Ahead: The Governance Imperative in a Data-Saturated World

The volume of data being generated globally is not going to decrease. The regulatory expectations governing how that data is used are not going to relax. The competitive pressure to generate faster, more accurate insights from data is not going to ease. What this means, practically, is that the organizations that invest in data governance now, that build the frameworks, the roles, the pipelines, and the culture that make data reliable, will compound that advantage over time, while those that defer the investment will find the gap between their capabilities and their competitors' capabilities widening.

Kanojiya's work across healthcare, enterprise software, and distribution represents a practical demonstration of what that investment looks like at the operational level. It is not glamorous work in the way that deploying a new AI model or launching a new analytics platform might be. It is foundational work, the kind that rarely gets credited in headlines but that determines whether everything built on top of it actually functions. As the data economy matures and organizations begin to understand more clearly why their analytics investments have underperformed, the practitioners who know how to build trustworthy foundations will become the most consequential people in the room.

The data exists. The trust does not. That is the problem. And it is a problem with a solution.

Sunil Kanojiya  ·  Data & Analytics Manager  ·  ABS Wholesale, California  ·  Engineering Trust in Data  ·  PMP®  ·  Advisory Board Member, IJIEEE / ITR / Science Guru