Document intelligence
A design for invoices, contracts, and business documents: field extraction, structured output, rule validation, and human review of exceptions.
SENXOS / LABS
Turn technical ideas into examples you can inspect.
Explore how AI, code and human review share responsibility in documents, data and workflows before considering further development.
Document amount validation
Interactive rule demonstration, not a customer case. Fields are predefined sample data. Arithmetic checks run in your browser; no OCR, model inference, uploads, or storage are involved.
Sample document fields
- System design service
- 1,250.00
- Interface integration service
- 750.00
- Subtotal
- 2,000.00
- Tax
- 120.00
- Document total
- 2,120.00
Sample currency: CNY. Calculations use integer minor units.
Rule checks
- Line items equal subtotalPassed
- Subtotal plus tax equals document totalPassed
Amount checks passed. Content and access still require workflow review; arithmetic does not establish semantic correctness.
Structured validation result
{
"sample": true,
"currency": "CNY",
"unit": "minor",
"lineTotal": 200000,
"expectedTotal": 212000,
"difference": 0,
"linesMatch": true,
"totalMatches": true,
"requiresReview": false
}Solution concept: Design concepts; online interaction is not yet available
Private knowledge retrieval
A design for internal policies and technical documents, connecting ingestion, retrieval, citations, and access controls so answers can be checked.
Workflow agents
Models identify intent and extract parameters; permission checks, rules, and human approval constrain tool calls and state changes.
Visual inspection
A design showing how image recognition, business thresholds, review queues, and result tracking can work together.
Questions, answered
Do demos upload or store data?
Current rule demos use prepared examples in your browser. They do not upload, store, perform OCR or model inference, or grant operational approval.
How do you handle incorrect AI answers?
Retain sources, use representative tests and rules, and route missing evidence, exceptions and important actions for human confirmation.
Why validate before development?
Prototypes and representative data help agree assumptions, usable results and scope before committing to the next stage.
Can we start without a full specification?
Yes. Describe the users, tools, problem and intended outcome. We can clarify whether a prototype and further development are appropriate.
Start your next project with a clear brief.
Share your goals, current workflow, and what needs to improve. We help clarify requirements, assess feasibility, and plan a suitable development approach.