Agentic AI Projects for ERP

Five model-agnostic, multi-agent solutions for Oracle ERP processes. Each project puts AI agents inside a real finance or procurement workflow. Agents investigate, reason and recommend; deterministic controls, human approvals and a complete audit trail govern every ERP action.

Reference builds in progress on an Oracle Fusion Cloud data set, each with a validation plan of 100+ scripted business scenarios

AP Invoice Processing & Exception Resolution

Oracle Fusion Payables · EBS AP

Moves clean supplier invoices from receipt to a validated Oracle invoice without manual keying, and investigates invoices on hold, resolving them or packaging them for a decision.
View Project
Financial Close & GL Variance Investigation

Oracle Fusion General Ledger · EBS GL

Investigates general-ledger variances from balances down to journals and subledger transactions, then prepares evidence-backed explanations and proposed adjusting journals for review.
View Project
Intelligent Procure-to-Pay: Request to Governed PO

Oracle Fusion Procurement

Turns a plain-language purchasing request into a structured requisition, checks policy and budget, compares approved suppliers and contracts, and prepares the requisition for approval.
View Project
AR Collections & Cash Application

Oracle Fusion Receivables · Collections

Prioritizes overdue accounts, investigates disputes, prepares collector-approved outreach and matches incoming receipts to open invoices.
View Project
ERP Security, Access & Segregation of Duties

Oracle Fusion Security · EBS Responsibilities

Assesses access requests against deterministic segregation-of-duties rules and actual user activity, and recommends least-privilege access with approval and audit controls.
View Project

Shared platform architecture

All five projects run on the same governed platform.

Business event or request
CrewAI orchestration
Specialized agents
MCP tools
Oracle ERP APIs and data + RAG policies + deterministic rules
Human approval
Controlled ERP action
Audit

Orchestration: CrewAI crews define role-based agents, tasks and hand-offs. Oracle tools attach through CrewAI MCP server adapters, so every agent calls the same governed tool layer.
Agent runtimes: Claude Agent SDK with Claude is the primary build and reasoning runtime. It uses subagents for deep investigation, MCP for tool access, and pre-tool-use hooks that block any ERP write lacking an approval token. The same agents run on Google ADK with Gemini and OpenAI Agents SDK with GPT through a common model-adapter interface, switched by configuration.
Oracle integration: A Python MCP server exposing typed read tools (Fusion Cloud REST resources, BI Publisher/OTBI reports, read-only SQL views for E-Business Suite) and a small set of write tools that execute only after human approval.
Grounding and control: RAG over policies, contracts and approval matrices in Oracle Database 23ai AI Vector Search; a deterministic rule engine for tolerances, thresholds and SoD; an immutable audit log of evidence, tool calls, model, approver and outcome.
Engineering: Python, Pydantic-validated agent outputs, a pytest scenario-evaluation harness comparing Claude, Gemini and GPT on accuracy, latency and cost, OpenTelemetry tracing and Docker, built in VS Code with Claude Code.

Portfolio coverage

Transactional automation (AP)
Financial reasoning and SQL analytics (Close)
End-to-end process orchestration (P2P)
Working capital (AR)
Governance and risk (SoD)

All five run on one model-agnostic, MCP-governed Oracle ERP platform.

AP Invoice Processing & Exception Resolution

Financial Close & GL Variance Investigation

Intelligent Procure-to-Pay: Request to Governed PO

AR Collections & Cash Application

ERP Security, Access & Segregation of Duties

Shared platform architecture

Portfolio coverage

Have an AI Automation Challenge?