ProcurAiQ classifies every line of spend, watches the commodity markets, models should-cost and predicts supplier risk on one Neo4j knowledge graph — then assembles the whole negotiation brief in minutes. The category manager keeps the decision; the AI does the work that used to take days across 35 spreadsheets.
Spend lives in ERPs, spreadsheets and email. By the time it's compiled it's stale — so the cheapest invoice often wins, even when it is the most expensive supplier on true cost.
The Amplinth loop, tuned to procurement — every answer cited back to the graph, every action human-checkpointed.
Ingest ERP transactions, supplier records, commodity indices, contracts & quality data onto a Neo4j graph.
Classify spend, forecast prices, predict supplier failure & trace risk with Neo4j GDS. No black box.
QATC-adjusted rankings, price-review-eligible contracts & a fully-assembled negotiation brief, with citations.
The category manager approves classifications, sets targets, selects suppliers and awards — gate-ruled.
Append-only audit of every AI step; QATC & savings ledger.
Category, Spend, Market, Supplier, Cost, Contract, Risk, Simulator, ESG, Quality, Decision and Reporting share one Neo4j substrate and one BFF — so its signature analytic, QATC, propagates across all of them.
The Category service runs a BERT-based classifier that maps every transaction to a multi-level taxonomy, while the Spend service lands a real-time spend cube and flags anomalies, duplicate invoices and off-contract maverick spend — human override always supported.
The Market service tracks commodity indices and price history, forecasts across 1/3/6/12-month horizons, and raises commodity alerts when a movement crosses a level that should trigger a contract or price review — with news, sentiment and regulatory signals per category.
The Cost service builds index-linked should-cost and clean-sheet models, flags quotes above fair price and decomposes the gap. BPID / QATC — Quality-Adjusted Total Cost — exposes the hidden downstream cost price-only tools can't see and re-weights every supplier scorecard around it.
The Supplier service computes 360° scorecards, scans financial health and ESG daily via D&B / Experian / EcoVadis connectors, and discovers qualified alternatives. Neo4j GDS — PageRank, centrality, shortest-path — traces how a shock propagates and flags single points of failure and concentration risk.
The insights service composes an RFQ skeleton, supplier letters, stakeholder memos and a full negotiation pack — assembling market evidence, cost breakdowns, alternatives and contract clauses into one brief, in minutes rather than days.
A planner with ~40 grounded handlers streams cited, confidence-tiered answers from the tenant's own graph. An ethics guard refuses unethical questions and a no-data guard returns a plain "not available" rather than inventing a number — every step on an append-only audit trail.
ProcurAiQ's BFF, ~20 services and 12+ micro-frontends are wired and running in the cloud today. The headline impact numbers below are targets or modelled worked cases, not realized client results — and are labelled as such.
The cheapest supplier on the invoice can be the most expensive on true cost.
Maturity: platform deployment is Realized (verified in code — services, BFF route table and micro-frontends are wired). Savings, cycle-time, classification accuracy and supplier-failure-warning figures are Target or Modelled until a cleared client baseline exists.
Connect ProcurAiQ to your ERP transactions, supplier records, commodity feeds and contracts. Customize the taxonomy, QATC hidden-cost dimensions and scorecard weights to your categories — the platform is reusable; the domain model is what we tailor. Operate the loop in production, where every classified transaction and approved award sharpens the graph.
QATC / Beyond-Price Intelligence prices suppliers on true total cost — a defensible, manufacturing-specific analytic that generic spend dashboards can't claim, propagated across all twelve modules.
A planner with ~40 grounded handlers, citations and confidence tiers, an ethics guard, a no-data guard that refuses to hallucinate numbers, and an append-only AI-step audit trail — governance made concrete in code.
Every classification, award and action is human-checkpointed and bounded by co-pilot gate rulesets and approval chains — the right posture for conservative, audit-heavy industrial buyers.
Book a demo on a real category — we'll show you the QATC ranking flip, the graph-native risk story, and a negotiation brief assembled in minutes, with the EVA case behind it.