Enterprises running SAP S/4HANA and Snowflake in parallel have largely solved the data movement problem. What they have not solved is the meaning problem across the enterprise.
When each platform maintains its own catalog in isolation, KPI definitions diverge, lineage stops at the platform boundary, and the data foundation required for reliable AI simply does not exist. The infrastructure is sound, but the governance layer is not.
The problem is not that these systems underperform. Each platform executes its function effectively. The gap is in what happens at the boundaries between those platforms, where catalogs do not connect, definitions do not travel, and lineage does not cross. In other words, the challenge is that they don’t always work together.
Each platform has its own catalog, language, and audience. What’s missing is a consistent way for users to navigate across all three and see how data moves from source to insight.
The Challenge: Fragmented Catalogs Across Platforms
The core challenge is not platform connectivity. SAP BDC for Snowflake solved data access issues at the zero-copy share level. Rather, the gap is at the catalog level, where business definitions, governance policies, and data lineage need to operate consistently across SAP S/4HANA, SAP Business Data Cloud, and Snowflake simultaneously.
SAP S/4HANA and BDC hold the operational truth of how the business runs. Snowflake Horizon adds the analytical power and governance needed to transform that data into decisions. Today, those catalogs are often managed in silos.
The result is a metadata landscape that creates three compounding problems: KPIs that cannot be traced to a single authoritative source, governance policies that do not travel with data as it crosses platform boundaries, and AI systems that cannot reason over business context they were never given access to. Each problem is manageable in isolation. Once together, however, they represent a structural barrier to enterprise AI.
A practical example: an organization running a global financial close process sources days sales outstanding (DSO) from SAP S/4HANA for operational reporting and from a Snowflake model for executive dashboards.
Both numbers are calculated from the same underlying transactions. They do not match. The difference is not a data error. It is a definition error, one that lives in the catalog layer, or more precisely, in the absence of one.
Without a shared catalog connecting the DSO definition in SAP to the logic applied in the Snowflake model, finance teams spend reconciliation cycles chasing a gap that governance should have prevented.
The Role of Snowflake Horizon in the Ecosystem
Snowflake Horizon provides the governance framework within Snowflake, including tagging, classification, and access policy management. Data lineage within Snowflake is surfaced through Access History and is extended through catalog integrations such as Atlan, which broaden visibility across pipeline stages.
Its focus remains on the analytical layer. It doesn’t naturally extend into SAP’s business semantics or the curated BDC extracts that bridge operations and analytics. This is where an enterprise-level marketplace, orchestrated through a system like Atlan, brings the pieces together.
Atlan as the Unifying Metadata Layer
Atlan operates above the native catalog layer. It ingests metadata from SAP S/4HANA, SAP Business Data Cloud, and Snowflake and centralizes it into a governed structure where lineage, definitions, and ownership are visible across the full data chain.
A finance analyst searching for invoice reconciliation data sees the relevant SAP source tables, the BDC data product that governs them, and the Snowflake model that feeds executive reporting, with lineage and accountability attached.
An engineer traces the same chain without leaving the catalog. What was previously siloed across three platforms becomes navigable in a single search.
Toward a Shared Enterprise Ontology
As this integrated catalog matures, it naturally enables the creation of a shared enterprise ontology—a common language that connects data across systems by meaning rather than syntax.
Concepts like Customer, Vendor, or Work Order become universal reference points, linking SAP’s operational depth with Snowflake’s analytical reach.
Agentic AI introduces a specific catalog requirement that does not exist in traditional analytics. When an AI system takes action based on data, whether generating a purchase order, flagging a compliance exception, or summarizing a supplier risk profile, the metadata chain behind that action needs to be complete and auditable.
An agent operating across SAP and Snowflake without a unified catalog has no basis for verifying that the data it is acting on has been governed consistently. The unified catalog is what makes agentic AI in the enterprise defensible.
From Catalogs to a Unified Data Marketplace
A unified catalog is not a tool selection decision. It is an architectural decision with consequences that extend across every reporting layer, every governance policy, and every AI system the enterprise deploys.
Each platform contributes what it is designed to deliver. Snowflake Horizon manages governance within Snowflake. SAP manages operational truth. SAP Business Data Cloud bridges the two with governed data products. Atlan makes the entire architecture legible, from the SAP source table to the executive dashboard, in a single catalog.
Together, they create a marketplace where data moves with meaning: governed, explainable, and ready for activation across every layer of the enterprise.
The organizations best positioned for enterprise AI are not necessarily the ones with the most advanced models. They are the ones with the most governed data. If you are unsure whether your current catalog architecture can support AI at the intersection of SAP and Snowflake, that is precisely the conversation IBM and Hakkoda are equipped to have with you.
Contact us to explore how IBM and Hakkoda can help you build a foundation for enterprise-scale decision intelligence and agentic AI.
You can also join us at SAP Sapphire in Orlando for an SAP-focused executive lunch co-hosted by IBM and Snowflake, taking place Wednesday, May 13th, from 1:00–2:00 PM in the IBM meeting space on the conference floor.
