AI-powered billing verification for construction (delivery notes + measurement = invoice)
I'm a construction site manager in landscaping (GaLaBau). Our industry software (greenGALA 8, BAUSU) is purely static – it generates output but performs no automated checks or reconciliations. Every verification is done manually: data entry, Excel cross-checks, triple manual control, handwritten delivery note sorting, and final invoice plausibility checks.
The core billing equation in construction is: Delivery notes + Measurement = Invoice. There is currently no tool that automates this reconciliation across the entire chain.
Three areas where AI support would create massive value:
Delivery note processing – Delivery notes are often barely legible: handwritten, dirty, incomplete, filled out on-site. Manual sorting and data entry takes unreasonable amounts of time. OCR-based extraction, even from poor-quality scans or photos, would save hours per project.
Measurement plan understanding – Measurements often exist as drawn plans or sketches, not digitally. If AI could read measurement plans and extract dimensions and quantities, that would eliminate another fully manual step.
Final invoice plausibility check – The most error-prone and time-consuming step. Reconciling delivery note quantities and measurements against billed positions. Finding deviations that can't be logically explained.
Concrete bug I already encountered: I uploaded a final invoice PDF + measurement list PDF to Lumo 2.0 Max (Thinking mode, web search enabled) and asked it to reconcile them. The processing log showed completion, but both times the output was: "Something went wrong." Reproducible twice with identical files and prompt.
Financial reality: I pay for Lumo out of my own pocket, not through a company license. Tax deductibility covers only a fraction. If Lumo can't handle the core use case – reliable PDF billing verification – it's financially uninteresting for me.
What I need: Reliable processing of two linked PDFs, structured table extraction, and automated discrepancy detection across delivery notes, measurements, and invoices – bundled in encrypted Lumo Projects. This applies to the entire construction industry, not just landscaping: civil engineering, road building, structural work, all trades.
Happy to provide sample documents or beta-test features.