Logistics / courier companies
The dispatcher takes a photo of the waybill, AI reads the parcels from it, and a floor worker scans barcodes and instantly sees what is missing and what is extra.
Courier and logistics companies receiving parcels against paper waybills.
Parcels were ticked off by hand on a printed list - slowly and with errors. A missing parcel only surfaced with a complaint, and numbers had to be retyped into the system.
A web app with two portals, AI data extraction from a photo, in-browser barcode scanning, list reconciliation, and export.
A photo of the parcel table goes into a Genkit flow with a Gemini model, which returns the driver, transport number, and shipment and parcel numbers in a fixed schema. When a model is overloaded, the app automatically falls back to the next one (2.5 Flash → 2.0 Flash → Flash-Lite). Workers scan barcodes with a phone camera or type the last five digits, and the list shows matched, missing, and extra parcels in real time. Lists and the activity log are stored in Firebase, and the result can be exported to CSV.
Receiving a delivery is a photo and a series of scans instead of ticking with a pen. Shortages are visible immediately on the floor, and data moves on without retyping.
LegalTech / law firms
A self-hosted law firm system: cases, documents with local OCR, paper archive, procedural deadlines, debt collection, and an AI assistant - without sending case files to the cloud.
~90k lines of code
Environmental services
An AI chatbot answering from the company knowledge base, plus a registry of issued and returned equipment. Customer support workload dropped by about 40%.
~40% less support workload
Customer service
AI reads customer messages from multiple channels, detects the topic, and drafts a reply in the company's tone - an employee just approves it. Response time dropped by about 70%.
~70% faster response time
