Document AI
Forge
An assistant that puts the historical document archive back to work, turning years of material into useful answers.
One foundation for AI, analytics and industrial automation.
One shared platform behind every service: access control, data storage, background processing and monitoring are already in place.
- Input
Files or data
A document, a technical drawing or an application data file comes in through the API.
- Platform
Common steps
Rules, permissions, storage and orchestration are handled once, for every case.
- Engine
AI / algorithm
The specialist engine only has to solve the vertical problem.
- Output
Operating result
A structured, traceable output, ready for a workflow or an operator interface.
- University of Trento
- Microsoft Azure
- Microsoft Graph
- Novotic
Shared capabilities behind every project
Every case reuses the same core capabilities instead of starting from scratch.
-
Data ingestion
We pull in files, documents and application data from APIs and company systems, including document archives and technical inputs.
-
Process orchestration
Background processing with work queues tracks state from start to finish, even through long or heavy processes.
-
AI & algorithms
Dedicated engines for retrieval, feature extraction, specialist models and decision support.
-
Security
Access control, separation between customers, and integration with managed identities, certificates and policy.
Five real applications on the same platform
Document search, technical drawing reading, decision support, demand forecasting and quality control.
Feature extraction
DataBlade Viewer
Reads DXF technical drawings and translates their geometry into data a piece of software can actually use.
Decision support
FlexoGPT
Generates the operating actions for blade straightening, from history and a model of the process.
Forecasting
Demand forecasting
A web application estimating future demand from historical sales data plus current and forecast market signals.
Quality control
Scalable quality control
Analyzes assembly curves, spots patterns and deviations, and returns a real-time quality assessment.
Where research meets the shop floor
The rigor of research, applied to real shop-floor problems.
Industry
Has to work on the shop floor
Solutions built to survive production reality, by people who start from your line and not from a manual.
Academia
Has to hold up to scrutiny
Methods developed with research partners such as the University of Trento and, where it matters, peer-reviewed.
The first step is an assessment on site
You tell us about your data, documents and processes; we show where AI can deliver a measurable result.