Executive Summary
A Fortune 500 industrial gas and specialty chemicals distributor, the largest US distributor of industrial, medical, and specialty gases, headquartered in Radnor, PA, with 18,000+ employees, 1,100+ branch locations in all 50 states, and $6.1B in annual revenue, was drowning in a manual purchase order process that no longer matched the scale or speed of its operations.
Thousands of monthly POs from hospitals, welding shops, food processing facilities, and electronics manufacturers were being processed by hand, creating a systemic 3–5 day approval cycle, a 10–30% error rate, and cascading delivery SLA failures.
After deploying the IQ Platform for intelligent PO ingestion and end-to-end workflow automation, the company reduced per-PO processing time from 18 minutes to under 3 minutes, cut the error rate below 2%, and recovered $1.4M in annualized operational costs, with full ROI in 4 months.
First-Year ROI
Reduction in Processing Time
Decrease in Error Rates
Reduction in Approval Cycle Time
SITUATION
The company operates as the backbone of America’s industrial gas supply chain, distributing oxygen, nitrogen, argon, acetylene, hydrogen, and specialty gas mixtures to an extraordinarily diverse customer base. Its buying segments include hospital procurement departments (medical-grade oxygen), automotive welding shops, semiconductor and electronics manufacturers (ultra-high-purity nitrogen), and food processing plants (CO₂ and cryogenic nitrogen for freezing). Each of these segments uses different ordering channels: large hospital networks submit EDI transactions, mid-market customers send Excel-formatted POs via email, small welding shops call in orders or email plain-text requests, and key accounts transact through a self-service web portal.
With 1,100+ branch locations managed from a centralized procurement operations hub in Radnor, PA, the company processes upwards of 8,000–12,000 purchase orders monthly. The ERP backbone, SAP S/4HANA, is deeply integrated with a warehouse management system (WMS) and a fleet scheduling platform. On paper, the procure-to-pay process was well-documented, with SOPs, vendor master files, and inventory procedures all formally maintained. In practice, however, the process was being executed almost entirely by hand.
TRIGGER
Two converging pressures made the status quo untenable in early 2024. First, post-pandemic demand recovery accelerated order volumes by approximately 22% year-over-year, straining a procurement team that had not grown proportionally. Second, a major electronics manufacturing customer, representing roughly $4.2M in annual contract value, formally cited two consecutive delivery SLA failures in a quarterly business review and issued a 90-day improvement notice, threatening to dual-source the account.
Leadership recognized that the bottleneck was not in the warehouse or the fleet, it was in the order intake office. A 3–5 day lag between PO receipt and confirmed delivery scheduling was making the company’s 48–72 hour delivery promise operationally impossible.
Simultaneously, SOX compliance audits flagged the manual approval routing process, specifically, the absence of a documented, timestamped audit trail for POs over $5,000, as a control deficiency requiring remediation within one fiscal quarter.
BARRIER
The root causes were well-understood internally, but the scale and variety of the problem made point fixes ineffective. Specific hurdles included:
Format Fragmentation at Scale: POs arrived as scanned PDFs, multi-tab Excel files, plain-text emails, and EDI 850 transactions, each requiring a different extraction approach. No single tool in the existing stack could handle all four.
18-Minute-Per-PO Manual Labor Cost: Each order required a specialist to open the attachment, manually key PO number, customer ID, product codes, quantities, and delivery address into SAP, then cross-check against the vendor master and product master. At 8,000+ POs/month, this consumed over 2,400 staff-hours monthly.
10–30% Error Rate with Downstream Consequences: Quantity mismatches, wrong product codes (e.g., industrial-grade vs. medical-grade oxygen), and duplicate order entries caused inventory allocation errors, mispicks, and customer-facing delivery failures. Resolving each error consumed an additional 25–40 minutes of rework.
SOX Control Gap: Manual email-based approvals left no structured, auditable approval log, a compliance exposure the CFO’s office could no longer accept.
Sequential Approval Bottleneck: Orders exceeding $5,000 required manager sign-off via email chain. With no parallel routing, a single manager’s absence could stall an entire queue. Average approval cycle: 3.7 days.
No Real-Time Inventory Linkage: By the time a PO reached the WMS for stock allocation, the data was often 4–6 hours stale, leading to over-promising on availability and back-order surprises at shipping.
Solution
The company deployed the IQ Platform (IQaaS), implemented by UTOFA, to build a fully automated, end-to-end PO intake pipeline, one that ingests every order format simultaneously, validates it against live ERP and inventory data, routes it for parallel approval, and confirms it back to the customer, all without a human touching the keyboard. Where the previous process required a specialist to manually bridge six disconnected systems, IQ’s Event-Driven Workflow Architecture orchestrates them as a single intelligent loop. The result is a procure-to-pay operation that runs at machine speed without sacrificing the human oversight that high-value and exception orders demand.
Inbox Monitoring & Ingestion: IQ Platform’s Stateful Event Handling System was configured to monitor three designated email inboxes and an EDI intake folder. Upon attachment detection, the system triggered the document processing pipeline automatically, handling PDF, Excel, plain-text, and EDI 850 formats in a unified queue.
Intelligent Data Extraction: The Intelligent Document Manager, powered by IQ’s SLM-First AI engine, was trained on 1,200 historical PO samples across all customer segments. It extracts PO number, customer ID, line-item product codes, quantities, unit pricing, delivery address, and requested delivery date with a validated accuracy of 98.4%. The SLM runs fully on-premise within the company’s VPC, no data leaves the corporate perimeter, supporting SOX data residency requirements.
Automated Validation & Vendor/Product Master Cross-Check: The IQ Platform’s Intelligent ETL engine performs three-way validation against the SAP vendor master (credit limits, blocked status), the product master (valid SKUs, grade classifications, a critical check for medical vs. industrial oxygen), and the WMS for real-time inventory availability. Mismatches are flagged and routed to a human review queue rather than silently failing.
Parallel Approval Workflow via Action Center: Orders exceeding $5,000 are routed simultaneously to the relevant regional manager via the IQ Action Center a structured, timestamped approval interface. This replaced the unstructured email chain with a SOX-compliant audit log. Approvals that are not acted on within 4 hours auto-escalate to a deputy approver, eliminating single-point bottlenecks.
SAP Auto-Entry & WMS Allocation: Validated and approved POs are written directly into SAP S/4HANA as confirmed sales orders via native REST API connectors, requiring zero IT support for screen-scraping or UI dependency. Inventory is allocated in the WMS in the same transaction cycle, locking stock within minutes of PO receipt.
Automated Customer Confirmation & Warehouse Notification: Upon successful SAP entry, IQ Platform’s workflow engine auto-generates a structured confirmation email to the customer (with confirmed delivery date) and a pick/pack task notification to the relevant branch warehouse, closing the loop end-to-end.
The Results
The IQ Platform implementation delivered transformational results by replacing manual, disconnected processes with intelligent orchestration:
Financial Impact: The company recovered $1.4M in annualized cost savings by eliminating 2,400+ manual processing hours/month, accounting for FTE reallocation and error rework cost reduction. ROI was achieved in 4 months, ahead of the 6-month internal target, driven by high-volume process velocity and 15 FTE offset.
Processing Excellence: Per-PO processing time dropped from 18 minutes to 2.4 minutes, an 87% reduction. Throughput scaled from 4–5 POs/hour to 18+ POs/hour per automated workflow instance, enabling the company to process 44% more volume without additional headcount.
Error Elimination: Error rate fell from 10–30% to 1.6%, driven by 98.4% extraction accuracy from the Intelligent Document Manager. Medical-grade/industrial-grade product code mismatches were eliminated entirely, a critical safety and regulatory achievement.
Approval Transformation: Approval cycle time dropped from 3.7 days to 19 hours through parallel routing via IQ Action Center with SOX-compliant audit trail. The CFO’s office cleared the control deficiency finding identified during SOX compliance audits.
Customer Relationship Stabilized: The electronics manufacturing customer relationship was preserved with sustained 48–72 hour delivery promise achievement. Customer satisfaction scores for order confirmation accuracy rose significantly, and delivery SLA failures became virtually non-existent.
Deployment Speed: Full production deployment was completed in 11 weeks, including SAP integration, document model training, and change management for 3 regional procurement teams.
"We were running one of the most complex order intake operations in the industrial gases sector, hospitals, fabs, welding shops, food plants, all with different PO formats and different urgency levels. The IQ Platform gave us a single intelligent layer that handles all of them. We stopped talking about backlog and started talking about throughput."
"The audit trail requirement alone would have justified the investment. We went from an email-chain approval process that our auditors flagged as a control deficiency, to a timestamped, structured workflow with full approver accountability, deployed in weeks, not quarters."
Summary
- Achieved 4-month ROI with 200% first-year return recovering $1.4M in annualized operational savings through elimination of 2,400+ manual processing hours monthly and error rework cost reduction
- Reduced per-PO processing time from 18 minutes to 2.4 minutes achieving 87% reduction and scaling throughput from 4–5 POs/hour to 18+ POs/hour per workflow instance
- Cut error rate from 10–30% to 1.6% through 98.4% extraction accuracy, completely eliminating medical-grade/industrial-grade product code mismatches critical for safety and compliance
- Decreased approval cycle from 3.7 days to 19 hours implementing parallel routing with SOX-compliant audit trail that cleared CFO control deficiency finding
- Preserved $4.2M annual contract value stabilizing electronics manufacturing customer relationship by achieving sustained 48–72 hour delivery promise and virtually eliminating SLA failures
- Deployed full production environment in 11 weeks including SAP integration, AI model training on 1,200 historical samples, and change management across 3 regional procurement teams
Key Metrics
| Key Performance Indicator | Before Automation | After Automation | Improvement |
|---|---|---|---|
| Processing Time | 18 minutes | 2.4 minutes | 87% reduction |
| Error Rate | 10-30 % | 1.6 % | 95% reduction |
| PO Throughput | 4-5 per hour | 18+ per hour | 300% increase |
| Monthly Manual Hours | 2,400+ hours | 300 hours | 87% reduction |
| Approval Cycle Time | 3.7 days | 19 hours | 79% reduction |
| Delivery SLA Achievement | 68 % | 96 % | 41% improvement |
