Concept card

SAP-RPT-1 / SAP-RPT-1.5 — Relational Foundation Model

As of 2026-07-24T15:00:00Z

Edited by Cédric-Olivier Mâry, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is SAP-RPT-1 / SAP-RPT-1.5 — Relational Foundation Model?

SAP-RPT-1/1.5 are research-stage foundation models pre-trained directly on relational table structure, not serialised text — specifically to understand that a column like KUNNR is a foreign key into KNA1, which LLMs treating tables as CSV/JSON entirely miss.

SAP-RPT-1 and its successor SAP-RPT-1.5 are SAP Research's foundation models purpose-built for enterprise tabular and relational data, rather than for text. The premise inverts the logic that made large language models useful for so many tasks: language models are pre-trained on enormous corpora of natural-language text, so that the statistical regularities of grammar, discourse, and factual association become reusable priors for almost any downstream language task. A relational foundation model applies the same idea to a different substrate — collections of enterprise database tables — betting that the structural regularities of relational data (foreign-key relationships, cardinality patterns, temporal ordering of transactional records, the way a header table relates to its line items) are rich enough to support the same kind of transfer learning.

What problem this actually solves

Why it matters

Named target tasks are concrete and enterprise-relevant: anomaly detection on financial line items, customer-master duplicate resolution, data-quality classification, missing-value imputation, and column-type inference.
The core limitation it addresses is well-documented: LLM serialisation of tables (CSV/markdown/JSON) loses join relationships and key semantics that RPT is designed to learn natively.
Honest maturity flag: as of mid-2025 these are explicitly research artefacts, not production-ready — set client expectations accordingly.

Key points

SAP-RPT-1 / SAP-RPT-1.5 — Relational Foundation Model is mastered only when it changes a named buyer decision.
Start with the semantic contract and control model before demonstrating the tool.
Use current SAP, analyst, study, KG, and news signals as evidence, not decoration.
Separate verified facts from directional trends and modeled assumptions.
Define owner, metric, threshold, support path, and rollback before scaling.
For AI use cases, measure reliability, cost, latency, safety, and human validation.
Leave a reusable operating asset: memo, checklist, control table, and exception log.
A premium answer is short, trade-off explicit, and defensible in a steering committee.

Terms used on this page

Decision owner: The accountable person who accepts the trade-off and funds the next action.
Semantic contract: The shared definition of business terms, metrics, entities, and access rules used by tools and teams.
Control plane: The layer that applies policy, access, lineage, monitoring, and escalation across the operating model.
Evidence grade: A label that separates verified fact, directional signal, modeled assumption, and field observation.
Adoption metric: The measurable behavior proving that the concept changed actual work after go-live.
Agent reliability: The consistency, cost, safety, and policy compliance of an agent across repeated runs.
Decision owner: The accountable person who accepts the trade-off and funds the next action.
Semantic contract: The shared definition of business terms, metrics, entities, and access rules used by tools and teams.

Sources

Full card available to members. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.

Open in the app →