Intelligent Procure-to-Pay: Request to Governed PO - Projects

The business problem

A manager writes, “We need 50 high-performance laptops for engineering within four weeks.” Before that becomes a purchase order, it has to be structured, checked against policy and budget, matched to approved suppliers and contracts, and approved by the right person.

Proposed solution

The crew turns this into a structured requisition, then checks policy and budget and compares approved suppliers and contracts. It recommends the best option with evidence and drafts the requisition in Oracle, committing nothing until the right approver signs off.

Workflow

Receive Purchase Request
Validate Policy and Budget
Compare Approved Suppliers and Contracts
Recommend Supplier with Evidence
Draft Requisition in Oracle
Route for Approval
Process Approved Requisition in Oracle
Record Audit Trail

Agents and responsibilities

Requirement: structured request, asking when details are missing.
Policy: RAG over purchasing policy, thresholds and IT standards, citing policy IDs.
Budget: cost center, funds, commitments.
Supplier & Contract: approved suppliers, blanket agreements, negotiated prices, delivery history.
Sourcing: ranks price, lead time, contract and risk, and rejects the cheapest supplier when it misses the need-by date.
Transaction: draft requisition, approval routing, PO tracking.

Oracle ERP integration

Oracle objects: Purchase requisitions, purchase orders, suppliers, blanket purchase agreements, receipts, budgetary control and the approval hierarchy.

Shared 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.

Technology stack

Project stack: Claude Agent SDK · CrewAI + MCP · Google ADK/Gemini · OpenAI Agents SDK/GPT · policy-driven model router

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.
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.

Controls and safeguards

Above-threshold purchases route to the correct approver.
Non-approved suppliers or insufficient budget stop the flow with reasons.
A policy-driven model router chooses Claude, Gemini or GPT per task using benchmarked accuracy, cost, latency and data-residency rules.
Grounding and control (shared platform): 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.

Validation plan

Scope: 100 scenarios over 50 suppliers, 20 contracts and 10 cost centers.

Targets

Reference build in progress. The figures below are validation targets for the build, not measured results.

100% Correct approval routing
0 Unauthorized transactions
Every Recommendation with evidence cited