How we replaced a dead-end legacy ERP and a sprawl of spreadsheets with one modern finance and manufacturing system, in about a third of the usual timeline.
Background
Airsan Corporation is a privately held industrial air filter manufacturer based in Milwaukee, Wisconsin, with more than 80 years in operation.
The business ran on MRP Plus, an on-premises ERP that had reached the end of its life through no fault of their own. The original developer had passed away, the support partner had retired, and the system sat on an aging database backend with no integrations. When the team last reached out for help, the remaining contact replied that he did not do installs. The system had begun producing a steady stream of errors, and because it connected to nothing, it could not support a modern workflow on its own.
To keep things running, nearly every function had grown a parallel layer of Excel. Finance built financial statements by hand from trial-balance exports, reconciled the bank in a daily tracking sheet, calculated use tax once a year in a separate file, prepared 1099s through an outside website, and keyed ACH payments manually at the bank. Manufacturing ran its production schedule, per-item costing, sub-assembly demand, work-order variance, and inventory counts in spreadsheets pulled out of the ERP and synced back by hand. As the controller put it, everything was triple data entry: quote, then order, then spreadsheet.
None of this was exotic, but replacing it is where most manufacturers stall. A full finance and manufacturing ERP implementation is traditionally a nine-month project, long and risky enough that aging systems tend to survive well past their usefulness. The challenge was to deliver the full scope without the full timeline.
What We Did
A three-day on-site assessment confirmed Microsoft Dynamics 365 Business Central as the right platform, at roughly 80% fit with very few gaps. From there we ran an AI-optimized implementation that compressed the active build into about three months, on schedule for a September 1, 2026 go-live, without cutting scope.
Full finance in Business Central. We restructured the chart of accounts to a standardized six-digit format with totaling and heading accounts, mapped posting groups to transaction types, built reusable journal templates for payroll, depreciation, and amortization, migrated three years of historical data, and turned on automatic inventory-to-GL posting, automatic 1099 generation, ACH and NACHA export, and use-tax handling through tax groups.
Full manufacturing in Business Central. We modeled multi-level bills of materials across roughly 1,300 finished goods and 600 raw materials, built granular multi-step routings (cut, crimp, oil, weld, spray, pack) to support their ISO 9001 documentation, and configured MRP with machine-calendar scheduling and automatic sub-assembly demand, standard costing with routing roll-up, native work-order variance, and scrap and yield tracking.
An AI-optimized delivery model. The speed came from how we ran the project, not from cutting corners. A Monday.com dashboard tracked project health, burn-down, and phases with AI, and our internal AI tools did the heavy lifting on the data: Claude standardized roughly 300 general ledger accounts in minutes instead of days, AI converted the roughly 2,224 legacy routing files (about 80% automated), and Granola captured every decision and process in real time as living documentation.
Claude and Copilot layered on. Business Central ships with Microsoft Copilot for natural-language queries against ERP data, so the team can ask for figures instead of digging through spreadsheets. Claude is already connected across their Microsoft suite through the custom plugins we built, with direct Business Central integration scoped as the immediate post-go-live step for cross-system work like family-level margin and substitution analysis.
Impact
The headline is speed with full scope. What is traditionally a nine-month implementation was delivered as roughly three months of active build, replacing a dead-end system and more than a dozen spreadsheets with a single modern ERP, on track for the September 1 go-live. That compression came entirely from an AI-optimized delivery model, not from narrowing the project.
The core outcome is structural: a business that ran on an unsupported system and a sprawl of spreadsheets now runs on one modern ERP, delivered in about a third of the usual time. Connecting Claude and Copilot directly to Business Central after go-live is projected to remove further manual analysis across finance, manufacturing, and sales, and the recurring hour and dollar savings will be quantified once the system reaches steady state.
- Financial statements come from native Business Central reporting instead of hand-built trial-balance exports.
- The monthly inventory-to-GL journal posts automatically instead of by hand.
- 1099s generate inside Business Central instead of through an outside site.
- Sub-assembly demand is calculated by MRP instead of tracked on spreadsheets.
- More than a dozen spreadsheet processes collapse into one platform.
Active build vs a traditional nine-month finance and manufacturing rollout
Roughly 1,300 finished goods and 600 raw materials modeled in Business Central
Legacy routing file conversion across roughly 2,224 files
Go-live target on schedule after full-scope delivery
“Everything used to be triple data entry: quote, then order, then a spreadsheet. Moving our finance and manufacturing into one system instead of a dozen spreadsheets changes how we work every day.”
Software: Microsoft Dynamics 365 Business Central, Microsoft Copilot, Claude, Monday.com, Granola, HubSpot, Microsoft 365
Similar Case Studies
More work from similar industries, or other recent stories from our library.

A full technical SEO overhaul in a third of the time
How we delivered a full technical SEO overhaul in about a third of the usual time using AI-native tools like Cursor, Semrush, and Claude.

Company-wide Claude enablement in about six weeks
How we took an 80-year-old manufacturer from zero AI use to a team building its own skills and automations in about six weeks.

Structured Copilot enablement for ~2,500 service staff across two license tiers
How we brought structured Copilot training to roughly 2,500 service staff across two license tiers, then layered discovery, a power-user cohort, and train-the-trainer to make it stick.
Tell us your biggest problem.
We'll show you the ROI.
Drop the problem in the box. A Wyecliff partner replies inside one business day with two ideas you can ship in 30 days. No pitch deck, no sales call required.
