Generative AI · SAP BTP · Design Thinking in Dearborn
How a major American automaker, SAP, and Mindset built an AI-powered FX reconciliation app and won the BTP Hackathon in its headquarters
Major American automaker Delivered 2025
the automaker, SAP, and Mindset ran a joint Design Thinking workshop and BTP Hackathon in its headquarters. The cross-company the automaker the hackathon team won with a currency-revaluation tool, an AI-assisted app that turns a reactive, multi-day month-end FX reconciliation into proactive, single-screen anomaly detection.
By the numbers
-
1st
Place at the the automaker SAP BTP Hackathon in its headquarters
-
500+
Company codes in scope for FX reconciliation
-
40 currencies
Covered by the a currency-revaluation tool prototype
-
1 screen
Replacing multiple S/4HANA transactions and Excel
Before
Reactive, multi-system month-end close
- FX reconciliation spread across multiple SAP transactions with no single report or drill-down view.
- Analysts pulled trial balances and exchange rates manually, stitching results in Excel across hundreds of company codes.
- Discrepancies surfaced only when the Profitability and Analysis team called, triggering late-night scrambles that delayed the financial close by days.
After
Proactive AI-driven anomaly detection on BTP
- a currency-revaluation tool proactively surfaces FX anomalies before the Profitability and Analysis team calls.
- A single screen replaces multi-transaction navigation, with AI-generated explanations for each mismatch.
- Validated target architecture on SAP BTP (Build, HANA Cloud, AI Core, Joule) ready for the automaker's production roadmap.
Why this matters
the automaker runs one of the most complex SAP estates in the world. A three-way Design Thinking and hackathon sprint with SAP and Mindset in its headquarters produced a winning BTP plus AI prototype that targets a real month-end close problem, and showed what the AppHaus model looks like when it runs at the automaker scale.
The challenge
Every month-end and quarter-end, the automaker's General Ledger analysts face a foreign-exchange revaluation grind. Reconciling valuation differences across hundreds of company codes and dozens of currencies means pulling trial balances and exchange rates from multiple SAP transactions and stitching them together in Excel. There is no single report and no drill-down view.
The work is reactive by design. Discrepancies surface only after the Profitability and Analysis team calls, at which point analysts work late into the night to reconcile and the financial close slips by days. The underlying data volume is enormous: roughly 57 million rows in ACDOCA for a single month.
the automaker wanted to know whether SAP BTP and generative AI could flip this workflow from reactive cleanup to proactive detection. The hackathon was the proof point.
What we did
the automaker, SAP, and Mindset Consulting ran a structured innovation sprint in its headquarters. It opened with a Design Thinking Workshop on October 6, 2025, where the joint team mapped the as-is analyst journey, voted on the highest-pain challenge (month-end and quarterly close review taking too long), and storyboarded a north-star vision: 'Efficiently Verifying Currency Revaluations.'
That brief fed the SAP BTP Hackathon on October 29, where the cross-company the automaker the hackathon team, drawing members from the automaker, SAP, and Mindset, built a currency-revaluation tool on SAP Business Technology Platform. The target architecture used SAP Build Apps, Process Automation, and Work Zone Standard Edition alongside SAP Business Application Studio, HANA Cloud, Cloud Identity Service, AI Launchpad, AI Core, and Joule.
The app proactively notifies analysts when FX anomalies appear, surfaces likely reasons for mismatches, and lets them drill straight from a summary view into a specific company code's postings on a single screen, replacing the current cycle of navigating multiple S/4HANA transactions and copying data into Excel.
The outcomes
The the automaker the hackathon team won the hackathon. The a currency-revaluation tool prototype showed the automaker's General Ledger and Enterprise Architecture teams a concrete BTP plus AI path off the spreadsheet treadmill and toward a faster, lower-risk month-end close.
Beyond the trophy, the three-way collaboration produced a validated target architecture and a working prototype that the automaker's teams can carry forward as a production roadmap item. The Mindset team demonstrated its role as the only US-based SAP AppHaus partner by co-leading the Design Thinking and BTP build alongside SAP and the automaker.
The prototype replaces a reactive, multi-system reconciliation workflow with proactive AI-driven anomaly detection and single-screen drill-down, directly targeting the analyst hours and close-delay risk the automaker identified in the Design Thinking session.
Recognition
- Winning team, Ford SAP BTP Hackathon, Dearborn, Michigan (October 2025)
If we built this today
Concept · not delivered scopeFrom all-nighter to one screen.
This is a forward-looking concept, not the scope we delivered on this engagement. It is the build we would reach for now, grounded in SAP that ships today.
the automaker's GL analysts hand-reconcile FX valuation differences across hundreds of company codes and dozens of currencies every month-end, and the gaps only surface after someone calls, so here is how that grind looks built on SAP Business AI today.
The data product
Working Capital Insights
Grounds the agent in governed close and valuation data from SAP Business Data Cloud, so anomaly scoring runs against one trusted view of ACDOCA, rates, and balances instead of stitched-together Excel exports. The Knowledge Graph gives it the real meaning of FCV, FCTA, and company-code structure.
Intelligent Application on SAP Business Data Cloud
The Joule agent
Currency Revaluation Reconciler
Reads ACDOCA postings, the FCV and FCTA valuation runs, exchange-rate tables, and trial balances across every company code and currency, then flags the valuation differences that look off and drafts the drill-down with a plain-language reason for each one. It hands the GL analyst a ranked anomaly list before the close, not after the Profitability team calls.
SAP S/4HANA, SAP AI Core, SAP HANA Cloud, SAP BTP, Joule · PROPOSE · Hours to reconcile valuation differences at month-end (currently ~30) and days added to the close
The Fiori app
Foreign Currency Valuation (with embedded Joule)
The standard S/4HANA foreign-currency valuation and period-end close apps now take a Joule prompt, so an analyst can ask why a company code's FCV moved and get the contributing postings and rates without hopping across transactions. Where a single named anomaly-review app does not exist yet, this is the GL close app family with Joule reading the same data.
Embedded in the Fiori launchpad
The honest first move is mining the real close process in SAP Signavio to see where the reconciliation hours actually go, mapping the BTP and S/4HANA estate in SAP LeanIX, then letting MIND accelerators carry the hackathon's a currency-revaluation tool logic from prototype into a grounded agent.
What we built
-
One agreed north-star challenge and solution direction feeding directly into the hackathon build
Design Thinking Workshop
A structured journey-mapping session with the automaker, SAP, and Mindset on October 6, 2025 in its headquarters. The team mapped the as-is analyst experience, voted on the highest-pain challenge, and storyboarded the target vision.
-
Hackathon-winning prototype targeting month-end close across hundreds of company codes and dozens of currencies
a currency-revaluation tool app
An AI-assisted SAP BTP application that proactively detects FX anomalies across hundreds of company codes and dozens of currencies, explains likely mismatch reasons, and replaces multi-transaction navigation with a single drill-down screen.
-
Full SAP-native BTP build, no off-platform components
SAP Build foundation
Target architecture built on SAP Build Apps, Process Automation, and Work Zone Standard Edition, with Business Application Studio as the development environment.
-
AI-driven anomaly explanation embedded in the reconciliation workflow
Generative AI layer
Integration of SAP AI Launchpad, AI Core, and Joule to surface anomaly explanations and analyst guidance directly inside the app, reducing the need to context-switch to external tools.
-
Architecture validated against roughly tens of millions of finance rows per month
HANA Cloud data layer
SAP HANA Cloud backing the app's data access layer, with SAP Cloud Identity Service for secure access, enabling queries across ACDOCA at the scale the automaker requires.
-
Winning team: the automaker the hackathon (the automaker + SAP + Mindset)
AppHaus-style cross-company delivery
Mindset led the Design Thinking facilitation and BTP build as the only US-based SAP AppHaus partner, working alongside SAP and the automaker engineers in a single integrated hackathon team.
A look at the work
