Customer story · Manufacturing
A manufacturing customer
Financial close analytics for a manufacturing customer
Journal-entry exports from several entities and ERP generations become one auditable close dataset, with KPIs and a bilingual assistant on top of a deterministic pipeline.
Executive summary
The situation
Shared-service finance closes across entities and ERP generations. Different export structures make it difficult to compare posting activity, especially around a calendar-defined closing window.
What was built
Nexus detects file types, unifies and enriches records, and computes close KPIs using the working-day calendar. Uploads supersede rather than mutate history, with a bilingual assistant for questions.
Why it matters
Gives finance a common analytical view with inspectable history. Core calculations remain deterministic; the assistant supports analysis rather than changing entries or the rules behind the figures.
- Primary users
- Shared-service finance staff
- Core systems/data
- Journal exports across entities and ERPs
- Human control point
- Immutable history; AI restricted to analysis
- Nexus role
- File detection, unified KPIs and versioning
The challenge
One close, many entities and ERPs.
Shared-service finance works across legal entities and ERP generations with different journal-export structures. Comparable questions about posting activity require a common data model, not merely a collection of uploaded files.
Manual versus automated postings, user and department activity, and work inside or outside the closing window need explicit definitions. The business working-day calendar and preserved upload history are essential context for examining those figures.
What we built
A unified, auditable close dataset.
Finance uploads raw journal-entry exports. Nexus detects file types from headers, then unifies and enriches the records. Close KPIs cover entries by entity, automated versus manual posting, activity inside and outside the closing window, and user and department views. The window is defined by the business working-day calendar, not a vague reporting label.
New uploads supersede earlier versions rather than mutating the historical data. A bilingual assistant supports natural-language questions above that auditable foundation. The core pipeline remains deterministic: AI helps analyse and question the dataset, but does not establish the close calculations or alter the journal records.
How the workflow runs
Step by step.
01
Upload
Finance uploads raw journal-entry exports. File types are identified from their headers.
02
Unify and enrich
Data from different ERPs is combined into one model. Close metrics are calculated using the working-day calendar.
03
Supersede, not overwrite
New uploads replace earlier versions while history is kept. The bilingual assistant answers questions over the data.
What the business team sees
Close KPIs for shared-service finance.
Entity, posting-type, user, department and closing-window views provide a common basis for examining the different exports. Finance can also ask the bilingual assistant questions. The rules and preserved upload history remain the basis for the figures being discussed.
- Journal entries by entity
- Automated versus manual posting
- Inside / outside closing window
- Posting user and department views
- Bilingual assistant
How Nexus works underneath
The calendar and the history are explicit rules.
Detection, unification, enrichment and KPI calculation are deterministic steps. Nexus applies the working-day calendar to the closing-window logic and preserves earlier uploads when new ones supersede them. AI operates above that foundation for analysis; it does not rewrite the records behind the reported figures.
Architecture
How the layers fit together.
Business team
Close analytics and assistant
KPIs by entity · posting views · questions
Nexus orchestration
- 01File detection
- 02Unify and enrich
- 03Close KPIs
- 04Working-day calendar
- 05Immutable versioning
Connected systems and data
- Journal-entry exports
- Multiple ERP generations
Human control and audit
Auditable by design.
Because data is never overwritten, finance can see which upload produced which figures. The assistant analyses data; it does not alter it.
What this changes
One analytical basis. Gives finance shared KPI views across entities and ERP generations.
A defined closing window. The working-day calendar governs classification inside or outside the window.
Earlier data stays inspectable. Superseding uploads preserve history instead of overwriting previous inputs.
Your process
Discuss a similar workflow
Tell us about the documents, systems and decisions involved. We will discuss where Nexus orchestration fits and where your team stays in control.