Business Flows · Knowledge Graph
AI use cases
Every week our AI Use Case Pioneer agent reads what vendors, analysts and practitioners publish about AI in operations. Each use case is mapped onto the Business Flows process model and rated by business potential and implementation effort. This page lets you explore the whole curated catalogue.
Curated content · Business Flows Knowledge Graph
This catalogue is drawn from the Business Flows Knowledge Graph, the curated library of end-to-end processes, SAP scope items, compliance obligations and AI accelerators behind our consulting work. Every use case links to the public source it was found in, so you can check it yourself. In the Business Flows Browser each one also shows its full description, the processes it accelerates and the regulations that apply to it.
Live from the knowledge graph · as of 10 Nov 2026 · 1,094 added in the last 30 days
5,390
AI use cases
10
AI technologies
8
process domains
499
public sources cited
222
linked to compliance obligations
Explore the catalogue
747 of 5,390 use cases match. Click a bar or a cell to filter.
Where to start: potential × effort
Business potential against implementation effort, as rated in the knowledge graph.
By AI technology
The technology that powers each use case.
By end-to-end process domain
Where in the Business Flows process model each use case applies.
A library that grows every week
Use cases added to the knowledge graph since the pipeline started.
Matching use casesnewest first
LLM · Finance
Generative AI Retail Customer Service Assistant
Potential: HighEffort: MediumSource: econsultancy.com (opens in a new tab) · added 19 Oct 2026
LLM · Finance
Governed AI Workflows with Prompt Injection Protection
Potential: MediumEffort: MediumSource: sumsub.com (opens in a new tab) · added 19 Oct 2026
LLM · Finance
Guided Buying with Conversational AI
Potential: HighEffort: LowSource: news.sap.com (opens in a new tab) · added 19 Oct 2026
RPA + AI · Finance
GxP-Aligned Electronic Approval Routing
Potential: MediumEffort: MediumSource: ezintegrations.ai (opens in a new tab) · added 19 Oct 2026
IDP + LLM · Finance
Intelligent Invoice Matching and Exception Handling
Potential: HighEffort: MediumSource: news.sap.com (opens in a new tab) · added 19 Oct 2026
Machine Learning · Finance
Intelligent Procurement Source Recommendation
Potential: MediumEffort: MediumSource: news.sap.com (opens in a new tab) · added 19 Oct 2026
IDP + LLM · Finance
Intelligent Tax Document Processing
Potential: HighEffort: MediumSource: salesforce.com (opens in a new tab) · added 19 Oct 2026
IDP + LLM · Finance
Multi-Channel Invoice Intake and Classification
Potential: HighEffort: LowSource: tungstenautomation.com (opens in a new tab) · added 19 Oct 2026
Optimization · Finance
Payment Timing Optimization for Early Discounts
Potential: MediumEffort: MediumSource: precoro.com (opens in a new tab) · added 19 Oct 2026
LLM · Finance
Policy-Driven Segregation of Duties Enforcement
Potential: HighEffort: MediumSource: mightybot.ai (opens in a new tab) · added 19 Oct 2026
IDP + LLM · Finance
Self-Improving Document Classification and Extraction
Potential: HighEffort: MediumSource: aws.amazon.com (opens in a new tab) · added 19 Oct 2026
LLM + Analytics · Finance
Spend Insights and Savings Opportunity Recommendations
Potential: HighEffort: MediumSource: news.sap.com (opens in a new tab) · added 19 Oct 2026
LLM + Analytics · Finance
Supply Chain Rebalancing Agent
Potential: HighEffort: HighSource: customertimes.com (opens in a new tab) · added 19 Oct 2026
IDP · Finance
Template-Free Invoice Data Extraction
Potential: HighEffort: MediumSource: tungstenautomation.com (opens in a new tab) · added 19 Oct 2026
Machine Learning · Finance
Three-Way Match Automation with Exception Detection
Potential: HighEffort: MediumSource: precoro.com (opens in a new tab) · added 19 Oct 2026
RPA + AI · Finance
Threshold-Based Invoice Approval Routing
Potential: MediumEffort: LowSource: mightybot.ai (opens in a new tab) · added 19 Oct 2026
IDP + LLM · Finance
Touchless Invoice Processing with Agentic AI
Potential: HighEffort: MediumSource: sumsub.com (opens in a new tab) · added 19 Oct 2026
LLM · Finance
Unified Approval Coordination Across Systems
Potential: HighEffort: MediumSource: leverx.com (opens in a new tab) · added 19 Oct 2026
Machine Learning · Finance
Vendor and CRO Performance Analytics
Potential: MediumEffort: MediumSource: appliedclinicaltrialsonline.com (opens in a new tab) · added 19 Oct 2026
LLM · Finance
Agent-Based Decision Support in Procurement
Potential: HighEffort: MediumSource: aws.amazon.com (opens in a new tab) · added 12 Oct 2026
LLM + Analytics · Finance
Agent-Based Procurement Sourcing Recommendations
Potential: HighEffort: MediumSource: news.sap.com (opens in a new tab) · added 12 Oct 2026
LLM · Finance
Agentic Code Review for Regulated Workloads
Potential: MediumEffort: MediumSource: aws.amazon.com (opens in a new tab) · added 12 Oct 2026
LLM · Finance
Agentic Exception Handling in Invoice Processing
Potential: HighEffort: MediumSource: aws.amazon.com (opens in a new tab) · added 12 Oct 2026
Machine Learning · Finance
AI Return Fraud Detection in Retail
Potential: HighEffort: MediumSource: cbsnews.com (opens in a new tab) · added 12 Oct 2026
Sourced, not generated
We don't invent use cases. Our agent extracts them from published material: vendor announcements, analyst research, case studies and the trade press. It records where each one was found. 5,365 of 5,390 entries link to a public source you can open and check. Each entry is then placed in the process model and rated for potential and effort.
- Found in a public source, with a link kept on every entry
- Mapped to the end-to-end process it accelerates
- Connected to the compliance obligations that apply
- Rated for business potential and implementation effort
Most-cited publishers in this selection
207 distinct sources in this selection.
The full picture is in the Business Flows Browser
Each use case comes with its description, the processes and SAP scope items it touches, the regulations that apply, and example tooling. It is all connected in one knowledge graph that you and your AI agents can query.
