Situation
Schupan is a leader in aluminum recycling, moving tens of millions of dollars in material every month. With this level of scale, timely and accurate reporting is critical. But their most important financial report — the Gross Margin Report (GMR) — was still being compiled manually across multiple spreadsheets.
It took their lead accountant 20+ hours each month to produce the report, which created a cumbersome invoicing cycle, slowed cash flow visibility, and tied up expertise that could have been directed toward high-value financial strategies. Meanwhile, logistics teams were buried in supplier emails and PDFs, re-entering data manually into the system — a process that introduced errors and slowed down operations.
Task
Schupan's leadership challenged Substantial to:
- Automate the Gross Margin Report so reporting could move from days to seconds.
- Streamline supplier logistics data by removing manual re-entry from emails and PDFs.
- Deliver tools that worked with Schupan's current systems while also laying a foundation for future AI-ready automations.
Action
Substantial embedded with Schupan's team, mapping real workflows step by step and learning the accounting language that often blocked past vendors. Instead of frameworks and slide decks, we delivered working solutions iteratively.
- Started with strategy, not tech. This engagement grew directly out of an AI Strategy Workshop we ran, where we identified high-value pain points and collaboratively narrowed in on finance and logistics as pilot use cases.
- Mapped the hidden workflow. Our team shadowed Schupan's accountants and operators, documenting every step of the manual reporting and data entry processes. This surfaced requirements that weren't written down anywhere.
- Built AI-ready automations on their existing stack. Instead of ripping out systems, we engineered lightweight automations inside Schupan's Azure environment, using modern ETL pipelines and structured data workflows to slot into their Excel- and ledger-based processes.
- Automated supplier data intake. We developed an AI-powered system that reads supplier PDFs/emails, normalizes formats, and feeds clean data directly into their database, reducing errors that previously cascaded downstream.
- Delivered fast iteration and real outcomes. By working side by side with Schupan's team in weekly demos, we tested assumptions, adapted to new requirements, and delivered a production-ready automation in weeks, not months.
Result
- 20 hours of monthly manual work eliminated.
- Reporting 8,000x faster: from 2.5 days to 10 seconds.
- Improved cash flow visibility with real-time numbers instead of week-old spreadsheets.
- Fewer errors in finance and logistics data.
- Experts freed to focus on strategic work like market hedging instead of spreadsheet assembly.
- Reusable automation patterns positioned Schupan for broader AI transformation.