AI Transparency & Disclosure

Asiyst provides AI-powered functionality that can process information supplied by users or applications and generate responses, recommendations, classifications, transformations, or other outputs.

Depending on the Service and configuration, Asiyst AI may be used for:

• Conversational interactions.
• AI-powered website assistants.
• Interactive AI avatars.
• Question answering.
• Knowledge retrieval and contextual responses.
• Processing information from configured knowledge sources.
• Content generation.
• Content transformation.
• Summarization.
• Classification and organization of information.
• Developer and application assistance.
• Automated workflows.
• Natural-language interaction with applications.
• Other AI-powered functionality made available by Asiyst.

The exact functionality available to a user depends on the Asiyst product, account configuration, enabled features, integrations, AI model, and applicable plan.

Not every feature of Asiyst necessarily uses artificial intelligence.

Some features may operate using conventional software, databases, rules, APIs, deterministic processing, or other technologies.

Where a feature uses AI to generate, transform, analyze, interpret, or respond to information, the applicable interface or documentation may identify the feature as AI-powered.

When you use an AI-powered feature, information may be processed through several components depending on the feature and configuration.

A typical request may involve:

1. A user or application submitting an input.
2. Asiyst receiving and processing the request.
3. Applicable authentication, authorization, security, and usage checks being performed.
4. Relevant instructions, configuration, or knowledge being applied.
5. The request being sent to an applicable AI model or provider where required.
6. The AI system generating an output.
7. Asiyst processing or formatting the output.
8. The resulting output being returned to the user or application.

The exact processing flow may vary between Services.

Depending on how you use Asiyst, AI processing may involve information such as:

• Prompts and messages.
• Questions submitted by users.
• AI conversation history where the relevant feature stores conversations.
• Information supplied through an application.
• Website content.
• Knowledge-base content.
• Documents or files submitted for processing.
• Product or service information.
• Developer-provided instructions.
• System or application configuration.
• Context supplied to an AI request.
• AI-generated responses.
• Technical information required to operate the request.
• Token and usage information.
• Error and diagnostic information where applicable.

The actual information processed depends on the specific feature and configuration.

Users and developers should not provide information to an AI feature unless they are authorized to submit and process that information.

Asiyst may use its own AI-related systems as well as third-party AI model providers.

Depending on the Service and configuration, third-party providers may include providers such as:

• OpenAI.
• Anthropic.
• Google.
• Other AI or machine-learning providers supported by Asiyst.

The providers used may change as Asiyst develops its Services, introduces new models, improves performance, manages availability, or adds additional integrations.

Where a third-party AI provider processes information, that provider may process the applicable request according to its contractual terms, technical documentation, privacy practices, and applicable data-processing terms.

Certain Asiyst Services may allow customers or developers to connect their own AI provider accounts or API credentials.

Where this functionality is available, the customer may select the applicable AI provider and model.

In these circumstances:

• The customer's request may be processed using the customer's selected provider.
• The applicable provider may receive information required to process the request.
• The provider's terms and policies may apply.
• The customer is responsible for configuring the integration appropriately.
• The customer is responsible for determining whether the selected provider is appropriate for the information being processed.
• The customer is responsible for complying with applicable laws and contractual requirements.

Customers should review the applicable provider's current terms, privacy policy, security documentation, and AI data-processing practices before connecting a provider to Asiyst.

AI-generated content is produced automatically using statistical and machine-learning systems.

AI-generated output may include:

• Text.
• Answers to questions.
• Summaries.
• Recommendations.
• Classifications.
• Generated instructions.
• Other machine-generated information.

AI-generated content may be inaccurate, incomplete, outdated, misleading, or inappropriate for a particular context.

The fact that information is generated by an AI system does not mean that the information is verified or factually correct.

You should independently review AI-generated content before relying on it.

Verification is particularly important when AI-generated content is used for:

• Legal matters.
• Medical or health-related matters.
• Financial decisions.
• Tax matters.
• Security decisions.
• Employment decisions.
• Compliance decisions.
• Safety-related decisions.
• Business-critical operations.
• Decisions involving significant financial or personal consequences.

Asiyst does not guarantee that AI-generated output will be accurate, complete, current, unbiased, or suitable for a particular purpose.

AI functionality provided through Asiyst is not a substitute for qualified professional advice.

Unless expressly stated otherwise, Asiyst AI should not be treated as:

• A lawyer.
• A doctor or medical professional.
• A financial advisor.
• An accountant.
• A cybersecurity professional.
• A licensed professional.
• A regulatory authority.

Where professional judgment is required, users should consult an appropriately qualified professional.

Asiyst distinguishes between processing information to provide the Services and using information to train general-purpose AI models.

As described in the AI Data Processing & Training Policy, Asiyst does not use customer content, prompts, conversations, uploaded information, knowledge sources, or other customer-provided content to train Asiyst's general-purpose AI models unless Asiyst expressly states otherwise or obtains the applicable permission.

This does not prevent Asiyst from processing information when necessary to:

• Provide requested AI functionality.
• Maintain the Services.
• Maintain security.
• Detect abuse.
• Prevent fraud.
• Troubleshoot technical issues.
• Monitor system performance.
• Comply with legal requirements.
• Provide customer-requested functionality.

Third-party AI providers may have separate policies regarding retention and model training.

When a third-party AI provider is used, the provider may have its own policies concerning whether API inputs, outputs, or other information are retained or used for model development or training.

Asiyst will not represent that a particular third-party provider does or does not use information for model training unless the applicable provider's current terms, documentation, or contractual arrangement supports that statement.

Users should review the applicable provider's current policies where this distinction is important to their use case.

Certain Asiyst processes may involve authorized human personnel where reasonably necessary for purposes such as:

• Customer support.
• Security investigations.
• Abuse investigations.
• Fraud prevention.
• Troubleshooting.
• Service reliability.
• Legal compliance.
• Incident response.

Not every AI interaction is reviewed by a human.

Asiyst does not represent that every AI-generated response has been reviewed, verified, or approved by a human.

Some Asiyst features may process requests or perform operations automatically without human review of each individual request.

Automated processing may be used to:

• Generate responses.
• Retrieve relevant information.
• Classify information.
• Process requests.
• Apply configured rules.
• Operate AI assistants.
• Support automated workflows.

Where applicable law provides users with specific rights concerning automated decision-making, those rights may apply depending on the nature and purpose of the processing.

Asiyst may provide AI assistants, interactive avatars, or similar interfaces that communicate with users.

Such assistants or avatars may:

• Answer questions.
• Navigate or explain application functionality.
• Retrieve information from configured knowledge sources.
• Provide contextual assistance.
• Guide users through supported workflows.
• Generate conversational responses.

The visual appearance, personality, name, voice, or presentation of an AI assistant does not mean that the assistant is a human.

Users should not interpret an AI avatar or assistant as a human representative unless the interaction is explicitly identified as involving a human.

Developers may use Asiyst to create AI-powered functionality for their own websites, applications, products, or services.

Developers are responsible for the behavior and configuration of their AI-powered applications, including:

• Instructions provided to AI systems.
• Knowledge sources.
• Data supplied to the AI system.
• User permissions.
• Third-party integrations.
• AI provider configuration.
• Application-level security.
• Applicable disclosures.
• Applicable privacy requirements.
• Applicable laws and regulations.

Developers should ensure that users are appropriately informed when they are interacting with AI where such disclosure is required or appropriate.

Where Asiyst allows customers to connect knowledge sources, AI systems may use information from those sources when generating responses.

Knowledge sources may include information provided through supported websites, documents, applications, databases, integrations, or other supported sources.

AI responses based on a knowledge source may still contain errors or omissions.

Customers are responsible for ensuring that they have appropriate rights and authorization to provide the relevant information to Asiyst and to use it as an AI knowledge source.

AI systems may process information originating from third-party sources.

The availability of information through an AI response does not necessarily mean that Asiyst independently verifies, endorses, or guarantees the accuracy of that information.

Users should evaluate information according to the requirements of their particular use case.

AI processing may involve personal information depending on how the Services are configured and used.

Users and customers are responsible for determining whether they have an appropriate legal basis, authorization, consent, notice, or other required permission to process personal information through an AI feature.

Additional information about personal-data processing is provided in the Asiyst Privacy Policy and AI Data Processing & Training Policy.

Users should exercise particular caution before submitting sensitive or confidential information to AI systems.

Unless expressly supported and appropriately configured for the applicable use case, users should avoid submitting:

• Authentication credentials.
• Passwords.
• Private encryption keys.
• Financial account credentials.
• Highly sensitive personal information.
• Confidential business information.
• Information belonging to another person without authorization.
• Information subject to contractual restrictions that prohibit such processing.

Users remain responsible for determining whether information is appropriate to submit to an AI service.

Asiyst uses reasonable technical and organizational measures designed to protect information processed through its Services.

However, no system, network, transmission method, or AI service can be guaranteed to be completely secure.

Customers are responsible for maintaining appropriate security controls for their accounts, applications, API keys, provider credentials, and connected systems.

Asiyst may use automated and other technical measures to detect, prevent, investigate, and respond to misuse of AI functionality.

These measures may include:

• Rate limiting.
• Abuse detection.
• Security monitoring.
• Request filtering.
• Account restrictions.
• API restrictions.
• Credential revocation.
• Feature restrictions.
• Investigation of suspected violations.

Asiyst may restrict or suspend AI functionality where reasonably necessary to protect users, third parties, infrastructure, or the Services.

AI functionality may depend on third-party infrastructure, model providers, networks, and other systems.

As a result, AI features may experience:

• Delays.
• Errors.
• Temporary unavailability.
• Rate limits.
• Model changes.
• Provider outages.
• Changes in response quality.
• Changes in supported functionality.

Asiyst does not guarantee uninterrupted availability of every AI model or AI-powered feature unless expressly provided under an applicable service agreement.

AI models and providers may change over time.

Asiyst may:

• Introduce new models.
• Remove or discontinue models.
• Change default models.
• Change model versions.
• Modify routing between supported providers.
• Adjust technical parameters.
• Change supported capabilities.

These changes may affect the behavior, speed, cost, or output of AI-powered features.

By using Asiyst AI functionality, you are responsible for:

• Reviewing AI output where appropriate.
• Using AI output lawfully.
• Protecting confidential information.
• Obtaining necessary permissions for submitted data.
• Ensuring appropriate user disclosures.
• Verifying important information.
• Maintaining appropriate security controls.
• Complying with applicable laws and third-party terms.

AI systems are probabilistic technologies and may produce different outputs for similar or identical inputs.

Asiyst does not guarantee that AI functionality will:

• Produce a particular result.
• Produce identical responses.
• Detect every error.
• Understand every request.
• Meet every user's expectations.
• Produce error-free output.
• Remain unchanged over time.

If you identify an AI-related issue, inaccurate behavior, security problem, harmful output, or other concern involving Asiyst, you may report it to Asiyst.

Reports should provide enough information to help Asiyst understand and investigate the issue, where it is safe and appropriate to provide that information.

Asiyst may update this AI Transparency & Disclosure as its AI functionality, technology providers, products, processing practices, or legal requirements change.

The current version published by Asiyst will describe the applicable AI practices at the time of use.

For questions regarding AI functionality, transparency, or disclosures, contact:

Asiyst Website: https://asiyst.com Email: support@asiyst.com