Home Insights Demos Public-safe demonstrations SAP AI Agent Demo Lab. Explore evidence-led agent patterns without live SAP connections, customer data, write-back, hidden prompts or autonomous enterprise decisions.
Start guided demo tour →
Featured demonstration
SAP Finance R2R Customization Optimizer
Purpose Explore how posting configuration, validations, substitutions, enhancements and linked custom developments may influence SAP Finance R2R posting.
Example input A non-confidential posting concern and one of four sanitized scenarios.
Evidence Mock configuration, customization and development evidence with explicit gaps.
Analysis Deterministic classification, risk, confidence and decision trace before optional narrative assistance.
Recommendation Advisory review candidates with evidence IDs and mandatory human confirmation.
Confidence Public demonstration confidence is capped at LOW.
Human-control boundary Finance, control and technical owners remain responsible for validation and decisions.
SAP write boundary No SAP connection and no write capability. Run public demo Agent demonstration
ST22 Dump RCA Agent
Purpose Explain ABAP runtime failures using approved dump evidence and architecture context.
Example input A sanitized short dump category and runtime symptoms.
Evidence Dump metadata, error context and approved diagnostic references.
Analysis Deterministic evidence extraction followed by governed explanation.
Recommendation RCA hypotheses and validation actions, not automatic fixes.
Confidence Depends on evidence completeness and remains advisory.
Human-control boundary An SAP expert validates the cause and remediation.
SAP write boundary Read-only demonstration; no live SAP access. Explore related architecture Agent demonstration
ATC Code Review Agent
Purpose Organize ATC findings and explain remediation considerations.
Example input Sanitized code-quality findings without source code or customer data.
Evidence Rule ID, priority, object category and approved finding text.
Analysis Finding classification, risk context and remediation options.
Recommendation Review guidance only; no automatic code change.
Confidence Limited by the supplied evidence.
Human-control boundary Developers and architects approve all code decisions.
SAP write boundary No code execution and no SAP write. Explore related architecture Agent demonstration
Batch Job Monitoring and RCA Agent
Purpose Explain job failures and performance signals using sanitized operational evidence.
Example input Mock job status, duration and error category.
Evidence Job metadata, log category and dependency status.
Analysis Failure classification and possible dependency correlation.
Recommendation Prioritized investigation actions.
Confidence Capped when logs or dependencies are missing.
Human-control boundary Operations validates evidence and recovery actions.
SAP write boundary No job restart, scheduling or SAP write. Explore related architecture Agent demonstration
Pre-deployment Check Agent
Purpose Organize technical readiness evidence before controlled deployment.
Example input Sanitized object scope, checks and dependency status.
Evidence Approved ATC, dependency and configuration-readiness signals.
Analysis Readiness gaps, risk and required confirmations.
Recommendation Go/no-go review inputs, not an autonomous approval.
Confidence Depends on check coverage and evidence freshness.
Human-control boundary Release and architecture owners make the decision.
SAP write boundary No transport release, import or configuration change. Explore related architecture Architecture boundary: All demonstrations use mock or sanitized evidence. No public demonstration connects to a customer SAP system, executes code, performs postings, changes configuration or approves an architecture decision.
