This Notice explains how Sento uses artificial intelligence and how data is handled when AI-enabled features are used. It supplements the International B2B Terms, International Privacy Policy, Public Data Processing Agreement, Subprocessor and Third-Party Recipient List, and Supported and Restricted Data Schedule.
1. What AI can do in Sento
Depending on Customer configuration, Sento may use AI to:
- analyze conversation and business context;
- retrieve information from Customer-approved knowledge sources;
- generate or suggest responses;
- classify, prioritize and route work;
- summarize conversations and documents;
- populate records and recommend next steps; and
- perform Customer-configured actions through integrations and APIs.
Customer controls the business purpose, connected systems, permissions, workflow rules and appropriate level of human review for its deployment.
2. Customer data is not generalized training data by default
Sento does not use Customer Content, Input or Output to train a generalized Sento model or a third-party generalized model by default.
Generalized training using Customer data requires a separate express authorization or written agreement and may occur only where the use is otherwise lawful, including any required legal basis, notices, rights, transfer safeguards, confidentiality restrictions and provider controls. A contractual opt-in alone does not override applicable privacy, intellectual-property or confidentiality law.
3. Customer-specific processing
Sento may index, create embeddings from, retrieve, classify and analyze Customer-approved documents, conversations, CRM context, databases and other sources to provide that Customer's Service. Sento may also use those sources for inference and Customer-specific configuration or adaptation where agreed.
These activities are Customer-specific service processing. They are not generalized model training merely because machine-learning or embedding technology is involved.
Customer-specific fine-tuning or other adaptation using Customer Content requires a separately agreed configuration, Order Form, statement of work or AI addendum and remains subject to the DPA and applicable privacy/security requirements.
4. Sento-operated and third-party AI components
Sento may use Sento-operated AI components and approved third-party model, inference or embedding providers. Where a third party processes Customer Personal Data on Sento's behalf, the provider must be covered by the applicable DPA, Subprocessor List and international-transfer mechanism.
A statement such as zero retention, no provider training, dedicated region or a specific processing location applies only where Sento expressly identifies that configuration in the current Subprocessor List, Order Form or other applicable agreement. Sento does not make those representations generically for every AI provider or model.
Sento is an AI-enabled application provider. Unless Sento expressly publishes otherwise for a specific Sento-developed model, this Notice does not represent that Sento is the provider of a general-purpose AI model.
5. Usage Data, diagnostics and Feedback
Sento may use Anonymous or Aggregated Data that no longer relates to an identified or identifiable person and cannot reasonably be linked back to Customer Content or a person to operate, secure, evaluate and improve the Service.
Usage Data that remains Personal Data is treated as Personal Data and is processed only for a disclosed purpose under an applicable legal basis and role.
Voluntary Feedback may be used to improve Sento. Feedback does not include Customer Confidential Information, Customer Content or Personal Data unless separately authorized.
6. AI Output
AI Output may be inaccurate, incomplete, non-unique, outdated or unsuitable for a particular decision. Customer is responsible for evaluating Output and configuring review, escalation or approval appropriate to the consequences of the workflow.
Sento does not promise that Output is copyrightable, unique or free of third-party rights. Rights in Output are governed by the Terms.
7. Transparency when a person interacts with AI
Where applicable law requires a person to be informed that they are interacting with an AI system, Sento will design or support the relevant directly interactive functionality so that the required disclosure can be presented clearly at or before the relevant interaction, subject to legally recognized exceptions.
For deployments subject to Article 50(1) of the EU AI Act, this includes enabling a natural person to be informed that they are interacting with an AI system unless that fact is obvious to a reasonably well-informed, observant and circumspect person in the circumstances.
Customer acts as the deployer or business operator of its configured workflow and must not intentionally disable a mandatory Sento disclosure mechanism. Customer remains responsible for any additional deployer-specific notice, sector-specific disclosure or local-language requirement that applies to its use.
8. Synthetic-content transparency
Where Sento provides functionality that falls within a mandatory machine-readable marking or detectability requirement for AI-generated or manipulated content, including Article 50(2) of the EU AI Act when applicable to the relevant functionality and application date, Sento will support the required technical transparency measures to the extent required by law and subject to legally recognized exceptions.
This section does not mean that every ordinary AI-assisted customer-support response must carry a visible label when applicable law does not require one.
9. Automated actions and human oversight
Sento may perform actions configured by Customer through connected systems. Customer is responsible for determining whether a workflow may materially affect a person's legal rights, eligibility, safety or access to important services and for applying the human review and safeguards required by applicable law.
Sento's ordinary support-response generation, classification and routing are not represented as systems intended to make eligibility, employment, credit, housing, insurance, healthcare, law-enforcement or similar high-impact determinations.
High-impact eligibility or legally significant decision use is not supported in the ordinary Sento Service. Such a category may become available only through a future separately documented product configuration after Sento updates the Supported and Restricted Data Schedule and implements the required legal, security and product controls.
10. Restricted AI uses and data
The Supported and Restricted Data Schedule identifies sensitive, biometric, health, government-identifier, children's and other regulated data, as well as high-impact automated decision use cases, that are prohibited in the ordinary Service.
A Customer must not intentionally configure a use or data category prohibited by that Schedule.
11. Regional rules
Mandatory regional requirements prevail over this Notice. Relevant privacy rights are described in the Regional Privacy Rights Notice, and international data transfers are governed by the International Data Transfer Addendum.
Questions about this Notice may be sent to [LEGAL_EMAIL].