Not used to train AI models
Synota does not use customer invoices, contract data, or system data to train external AI models or shared datasets.
How Synota Uses AI.
Synota uses automated analysis tools, including language models, to help interpret invoice documents and identify potential discrepancies.
These tools operate within a structured system designed for reliability and control.
- +AI is used to interpret and organize data
- +Synota performs all financial calculations and reconciliation logic
- +Customer data is not used to train external models
- +Customers remain in control and review all outputs
What AI does — and doesn't.
AI helps with
- +Extracting invoice data from PDFs
- +Categorizing charges into standardized billing groups
- +Interpreting billing structures — tariffs, taxes, regulatory charges
- +Summarizing invoice and reconciliation differences
AI does not
- ×Perform financial calculations
- ×Determine settlement amounts
- ×Modify contracts or billing terms
- ×Execute transactions or system changes
All calculations and settlement logic are performed directly within Synota's platform.
AI interprets. Synota calculates. You review.
Invoice reconciliation is not a single AI request — it is a multi-step process, and AI touches only the interpretation steps.
- 1AI assists
Invoice data is extracted and structured
- 2AI assists
Charges are categorized and interpreted
- 3Deterministic
Contract and reconciliation data are retrieved
- 4Deterministic
Deterministic calculations are performed
- 5You review
Results are summarized for review
This approach improves accuracy, transparency, and auditability.
Accuracy & review
Synota's automated analysis is designed to assist invoice review, not replace it. Because utility invoices vary widely, some classifications may require validation. Users should review results alongside original invoices, contract terms, and reconciliation outputs.
Your data stays yours.
Not stored by AI systems
Customer data is processed only to generate results. All documents and outputs remain within Synota's systems.
Reference data provides context
Synota may use structured reference data — utility tariffs, tax rules, regulatory charge definitions, and other trusted industry sources — to help interpret invoice charges. All reconciliation calculations are still performed within Synota.
Chat & interactive analysis.
What the chat feature does
The “Ask Synota” feature allows users to:
- +Ask questions about invoices and charges
- +Explore reconciliation results
- +Receive summaries and explanations
What it does not do
The chat feature is read-only. It cannot:
- ×Modify contracts
- ×Change calculations
- ×Affect settlement outcomes
- ×Update system data
It is designed only to help users understand existing information.
Optional by design
AI-assisted features are optional. Core platform capabilities — contract management, reconciliation calculations, and settlement workflows — operate independently of automated analysis tools, and customers may choose not to use them.
Reliability, safeguards & security.
Reliability & testing
- AI is limited to interpretation tasks
- Calculations are deterministic
- Outputs are tested against known invoice scenarios
- Results are reviewable and auditable
System safeguards
- Automated features do not execute system changes
- Core financial logic is isolated from AI components
- Outputs are stored and traceable
- Users remain responsible for final review
Security & compliance
Synota's use of automated analysis tools follows the same security and data handling standards applied across the platform — handled within Synota's controlled infrastructure, with external services used only for real-time processing.
Human review, by design.
Synota's automated analysis tools are designed to support — not replace — human review. Customers remain in control of how results are used and should validate outputs alongside source documents and contracts.
Talk to Synota