Fully autonomous project. The Procurement Codex is built, verified, and published end-to-end without manual authoring. Its core logic — the spine, the layer model, and the platform comparison rubric — is rebuilt and improved on every iteration for continuous method validation. Content is generated programmatically and refined each cycle: treat it as a directional learning aid, verify against primary sources, and send corrections — accuracy and fairness compound with each pass.
Episode 01 · The front door
Everything in procurement starts here — the moment a need appears and someone has to buy something. Get intake right and the entire cycle runs clean. Get it wrong and you spend the next fourteen steps firefighting maverick spend, rework, and shadow buying.
Below: the plain concept → how every platform does it (fair A vs B vs C) → then we stack process mining, AI, orchestration and people on top until you see why standalone is never enough.
The path from a raw need ("I need 50 laptops") to a structured, routed, policy-checked request that's ready to become a catalog order or a sourcing event. It is procurement's demand capture layer.
To capture demand at the source, route it to the right process, and enforce policy and risk checks early — before money leaks out through off-contract, maverick, or shadow buying.
One front door · no-wrong-door guided intake · auto-routing by category & risk · policy-as-code checks up front · high self-service adoption · short intake cycle time.
Same rubric for every vendor, 1–5. We state explicitly what each is best and worst at. Toggle platforms to compare.
| Platform | Best at | Watch-out | |
|---|---|---|---|
Scores are directional teaching aids based on typical deployments, not vendor benchmarks. Your mileage varies by configuration, module licensing, and integration maturity.
Stack layers onto the plain intake process and watch the architecture — and the outcome metrics — change. This is the whole thesis of the Codex in one control.
A real best-of-breed intake architecture — no single vendor owns all of it. The value lives in the seams.
flowchart LR R([Requester: natural language need]) --> AI[L4 GenAI intake copilot
classify + structure] AI --> O{L5 Orchestration
Zip / ORO — no-wrong-door} O --> LEG[Legal review] O --> SEC[IT security] O --> FIN[Finance / budget] O --> RISK[Third-party risk] LEG & SEC & FIN & RISK --> P[L1 P2P suite
Ariba / Coupa / Oracle] P --> ERP[(ERP · S/4HANA)] P --> PM[L3 Process mining
Celonis / PM4Py] PM -.conformance & bottlenecks.-> O
Standalone, the P2P suite handles requisitions but bounces approvals serially and can't see its own bottlenecks. Add orchestration and the four reviews run in parallel. Add a GenAI copilot and the requester never sees a form. Add process mining and the whole loop gets continuously tuned from its own event log. That is the synergy — and no single vendor sells all four honestly.
Pick one from each column. The Codex assembles the composite and calls out where the seams need engineering. Shown once here for intake; the same engine powers every episode.
Turn a raw need into a structured, routed, policy-checked request — before money leaks.
One front door · guided routing · policy-as-code · deflect to catalog · aggregate demand.
Multiple entry points · email/Slack requests · approvals in series · no category tag at source.
Zip/ORO lead on no-wrong-door intake; Coupa on UX; Ariba on network depth; Oracle on ERP-native; GEP/Ivalua on configurability.
L3 mining finds the bounce · L4 AI kills the form · L5 orchestration parallelizes reviews · L6 clarifies who owns what.
No single vendor does intake + orchestration + AI + mining well. Value is in the seams.