Capabilities
What it answers
Eight prompt categories cover the most common practitioner questions — the kind that previously meant a Slack message to a colleague or a 45-minute PDF hunt.
📋
SAP Incentive Management
Core platform concepts, pipeline logic, compensation plan structure, and
how the ICM engine works end-to-end.
🗄️
SAP HANA
HANA cloud vs. database distinctions, SAP IM's usage of HANA, and how it
replaced the Oracle stack.
🔄
CallidusCloud / Callidus
Legacy CallidusCloud Commissions context, Oracle-era architecture, and
the migration path to SAP IM.
📊
Data Dictionary
Key database tables, column definitions, and the TCMP schema — without
manually opening documentation.
⚙️
SAP Advanced Workflow
Workflow types, config patterns, routing logic, Groovy scripting hooks,
and approval escalation design.
🎓
Certificate Prep
Exam-focused Q&A, topic breakdowns, and scenario-based prep for SAP
Incentive Management certification.
💬
Interview Questions
Role-level interview questions for SAP IM practitioners — from beginner
fundamentals to senior implementation scenarios.
How It Was Built
The build
Built on Claude AI Agent — not a generic chatbot wrapper. Grounded on domain-specific SAP IM content so it answers like someone who's actually implemented the platform, not like someone who skimmed the help portal.
Layer 01 · Foundation
Claude AI Agent
The agent is built on Claude AI Agent — a low-code agent builder inside the
Microsoft Power Platform. It handles conversation orchestration, prompt routing, and the integration surface
into Microsoft 365 environments.
Layer 02 · Knowledge
Grounded on SAP IM domain data
The agent is grounded on curated SAP IM content — not just help documentation,
but practitioner-level material covering the TCMP schema, pipeline logic, Groovy scripting patterns,
compensation plan design, Advanced Workflow config, and the CallidusCloud-to-SAP-IM migration story. This is
what separates it from a generic LLM answer.
Layer 03 · Experience
8 guided prompt starters
The agent surfaces eight structured starting points that cover the most common
practitioner questions. Each starter is scoped to reduce hallucination risk and push the model toward the
grounded knowledge rather than general training data.
Tech Stack
⚡ Claude AI Agent
🔷 Microsoft Power Platform
🧠 Grounded Knowledge Base
💬 Conversational AI