AI Assistant Disclosure & Usage Policy

1. What Is Yumi?

Yumi is KCO’s artificial intelligence business assistant.

She is designed to assist with business communication, service information, enquiry handling and approved operational activities.

Yumi is an AI system, not a human employee.

Although she may communicate conversationally, use natural language and perform assigned business functions, her responses are generated or processed through software systems.

References to Yumi as a business assistant or operational director describe her assigned responsibilities within KCO. They do not establish independent legal personality, employment status or unrestricted decision-making authority.

KCO remains responsible for establishing the permissions, procedures and human oversight appropriate to its AI operations.

2. How Visitors May Interact with Yumi

Visitors may communicate with Yumi through approved website features and business communication channels.

Depending on the available functionality, Yumi may:

• Explain KCO’s website and digital services.
• Answer general questions about service options.
• Discuss prospective website projects.
• Collect information needed to assess enquiries.
• Explain quotation and development processes.
• Provide preliminary estimates based on stated assumptions.
• Assist with support and maintenance enquiries.
• Route matters requiring additional review.
• Support approved follow-up communications.

The functions available to a visitor may differ according to the communication channel, operational configuration and relevant permissions.

Not every capability described in this policy is necessarily available in every AI interaction.

3. AI Transparency

KCO aims to make the use of artificial intelligence clear to visitors and clients.

Yumi is presented as an AI assistant rather than a human representative.

Visitors should understand that responses may be generated using artificial intelligence and may be informed by approved business information, operational records and relevant conversation context.

AI-generated communications may sometimes resemble ordinary human business correspondence. This does not change the automated nature of the underlying assistance.

Where required by applicable law, KCO will provide appropriate information about the use of AI-generated content and AI interaction.

4. Accuracy and Limitations

AI systems can produce useful responses, but they are not infallible.

Yumi may occasionally provide information that is incomplete, outdated, misunderstood or incorrect.

Her responses should therefore be considered in the context of the question, available information and relevant business arrangements.

Visitors should seek confirmation before relying on AI-generated statements concerning significant financial commitments, legal matters, contractual obligations or technically consequential decisions.

KCO aims to maintain appropriate business knowledge and operational safeguards, but cannot guarantee the accuracy of every AI-generated response.

Where an error is identified, visitors are encouraged to bring it to KCO’s attention for review and correction where appropriate.

5. Quotations and Commercial Authority

Yumi may assist prospective clients by discussing requirements, explaining service options and preparing preliminary quotations or estimates within approved business rules.

Where sufficient information is unavailable, an estimate may rely on clearly stated assumptions.

An indicative estimate does not automatically constitute a binding commercial offer.

Final pricing, project acceptance and contractual commitments are determined by KCO’s authorised commercial procedures and applicable agreements.

Yumi must not be treated as having unrestricted authority to waive charges, create unsupported discounts, amend contractual obligations or approve exceptional arrangements.

Only commitments made through an authorised process can establish the relevant contractual obligations, subject to applicable law.

For further information, review our Website Development Terms of Service and Refunds, Cancellations & Payments Policy.

6. Human Oversight and Escalation

KCO establishes human oversight for matters requiring judgement, exceptional authority or further investigation.

Depending on the circumstances, these may include:

• Contractual disputes and unusual commercial requests.
• Requests involving substantial financial consequences.
• Complex complaints or contested decisions.
• Privacy and personal-data rights requests.
• Security incidents or suspected misuse.
• Requests outside Yumi’s assigned permissions.
• Matters requiring specialist professional judgement.

Yumi may help collect information and coordinate these requests, but human review or approval may be required before consequential action is taken.

Visitors may request human assistance through the applicable contact channels.

KCO does not represent that every AI action receives individual human approval before execution; oversight is determined by the relevant system permissions, risks and operating procedures.

7. Information Submitted to Yumi

Visitors may provide personal information when communicating with Yumi, including names, contact information, project requirements, business details and the content of enquiries.

Relevant information may be used to:

• Understand and respond to enquiries.
• Maintain continuity of business conversations.
• Prepare estimates and project assessments.
• Coordinate legitimate follow-up communications.
• Support operational record-keeping.
• Investigate errors, security concerns or service problems.

Visitors should provide information relevant to their enquiry and avoid disclosing unnecessary sensitive personal information.

Passwords, payment-card details, authentication codes and other confidential credentials should not be entered into ordinary AI chat messages.

Where secure access information is genuinely required, KCO should establish an appropriate protected method for supplying it.

8. Privacy and Personal Data

Personal information processed through Yumi is subject to applicable data protection requirements, including the General Data Protection Regulation (GDPR) where applicable.

The purposes and legal bases for processing depend on the nature of the interaction and the relevant business relationship.

Depending on the arrangement, KCO may act as a data controller or process information on behalf of a client under an appropriate agreement.

AI conversations and related operational records may be retained where reasonably necessary for legitimate business purposes, applicable legal obligations and agreed services, subject to relevant retention arrangements.

Where external technology providers process personal information on KCO’s behalf, appropriate contractual and data protection safeguards must apply.

For information about personal data, individual rights, recipients and retention, consult our Privacy Policy.

9. AI Providers and Technology

AI services may rely on language models, software components, infrastructure providers and other supporting technologies.

These systems may change as the technology develops and as operational requirements evolve.

KCO is responsible for evaluating the suitability of the technologies it selects and implementing relevant safeguards according to their intended use.

The involvement of an external AI provider does not automatically mean that all conversation data is publicly available or used to train general-purpose AI models.

Specific data handling, retention and training arrangements depend on the provider, service configuration and contractual conditions.

KCO’s published privacy information and applicable data-processing arrangements must accurately reflect the technologies and practices used.

10. Automated Processing and Decisions

Yumi may perform automated processing to classify enquiries, retrieve information, prepare responses, update authorised operational records and assist business workflows.

Such processing does not automatically constitute a legally significant automated decision under data protection law.

Where processing could produce legal or similarly significant effects on an individual, KCO must assess the applicable legal requirements and provide appropriate safeguards.

These may include meaningful human involvement, relevant information about the processing and the ability to exercise applicable data protection rights.

KCO does not authorise Yumi to make unrestricted high-impact decisions merely because a technical capability exists.

11. Responsible and Acceptable Use

Visitors and clients must use KCO’s AI services lawfully and responsibly.

Prohibited activities include:

• Attempting to obtain confidential information without authorisation.
• Seeking access to other clients’ data, records or systems.
• Submitting malicious instructions intended to bypass security controls.
• Using AI features to facilitate fraud, impersonation or unlawful conduct.
• Generating abusive, threatening or unlawful communications through KCO systems.
• Interfering with the availability or integrity of AI infrastructure.
• Attempting to manipulate system permissions or operational safeguards.

KCO may investigate suspected misuse and take proportionate protective action, including restricting access where appropriate and lawful.

General website usage rules are explained in our Website Terms of Use.

12. Managed Business AI for Clients

In addition to Yumi’s work for KCO, we offer Managed Business AI as a separate commercial service.

These solutions are intended to operate with approved KCO-developed websites and their managed infrastructure.

Managed Business AI is not offered as an unrestricted, standalone chatbot installation for arbitrary external websites.

Each client implementation requires an approved service scope, appropriate technical environment and applicable commercial agreement.

Available capabilities depend on the purchased service arrangement and may include business enquiries, approved workflows, knowledge retrieval and authorised operational assistance.

A client’s AI service is configured for its own business requirements and must not be assumed to possess the same permissions, integrations or operational responsibilities as Yumi.

13. Client AI Permissions and Boundaries

Client AI systems operate according to the permissions, integrations and business rules approved for their individual deployment.

The existence of a technical integration does not automatically authorise every action that integration could potentially perform.

Appropriate controls may be used to restrict:

• Access to confidential records.
• Financial or contractual commitments.
• Modification of operational data.
• Customer communication workflows.
• Access to third-party services.
• Actions requiring explicit human approval.

Clients are responsible for providing accurate operational instructions and identifying decisions that must remain under human control.

Material changes to permissions or capabilities may require technical assessment and additional approval.

14. Client Data Separation and Security

KCO aims to maintain appropriate separation between independently managed client environments and their business information.

Client AI implementations must not knowingly expose another client’s confidential data through unauthorised access or shared permissions.

Technical safeguards may include access controls, isolated infrastructure, restricted integrations and appropriate operational monitoring.

No AI system or digital infrastructure can be guaranteed entirely free from vulnerabilities or security incidents.

Suspected security issues should be reported promptly through approved communication channels.

Hosting and related responsibilities are addressed in our Hosting, Maintenance & Support Terms.

15. Knowledge Sources and Business Instructions

An AI assistant’s usefulness depends partly on the accuracy, completeness and relevance of the information available to it.

KCO and its clients may provide approved business information, procedures, service descriptions and operational instructions for relevant AI systems.

Clients are responsible for identifying errors in their supplied information and notifying KCO of material changes.

KCO may assist with approved knowledge updates according to the purchased service plan.

Updating reference material does not guarantee that every future AI response will be free from mistakes.

Material knowledge changes, specialist training, system restructuring and additional integrations may require separate work and charges.

16. AI-Generated Content and Ownership

AI-generated text and other outputs may be subject to applicable intellectual property rules and the terms governing the underlying technology.

Not all AI-generated material necessarily qualifies for exclusive copyright protection.

Clients must review AI-generated content before relying on it for legally sensitive, regulated or significant commercial purposes.

Where appropriate, clients should verify factual accuracy, originality, third-party rights and suitability for publication.

Ownership or licensing of commissioned AI-related deliverables is determined by the applicable agreement and relevant law.

KCO does not guarantee that AI-generated material is unique or free from similarity to content generated elsewhere.

17. AI Availability and Service Continuity

AI services depend on supporting infrastructure, software, integrations and potentially external model providers.

Service performance and availability may be affected by maintenance, provider incidents, connectivity problems, usage limits or technical failures.

KCO may perform necessary updates or temporarily restrict particular capabilities to protect security and operational reliability.

Specific availability commitments, usage allowances and support obligations are determined by the applicable service agreement.

No unconditional guarantee of uninterrupted AI operation is provided.

18. Monitoring and Quality Improvement

KCO may review authorised operational information to identify technical faults, improve the reliability of its services, maintain business knowledge and investigate security concerns.

Such activities must respect applicable privacy requirements and the roles and permissions established for the relevant system.

Monitoring and evaluation may involve human review where appropriate.

Quality improvement does not imply unrestricted permission to reuse client information or personal data for unrelated purposes.

Any use of information for AI model training or other distinct purposes must have an appropriate legal basis and be accurately disclosed where required.

19. Changes to AI Capabilities

AI technologies evolve, and KCO may update models, infrastructure, safeguards and supported features where reasonably necessary.

Changes may be introduced to address reliability, compatibility, security, provider requirements or service improvements.

Material changes affecting a client’s agreed service obligations are governed by the applicable contract and law.

New functionality is not automatically included in an existing subscription merely because it becomes technically available.

Additional capabilities may require assessment, implementation work and commercial approval.

20. Suspension and Termination of AI Services

AI services may be restricted, suspended or terminated where permitted by the applicable agreement and law.

Relevant circumstances may include serious misuse, security risks, material contractual breaches, unpaid charges or withdrawal of essential third-party technology.

Where reasonably practicable, KCO will communicate relevant restrictions and available steps for resolution.

Client data export, retention, deletion and technical shutdown arrangements must follow the applicable agreement and data protection requirements.

Additional payment and cancellation information appears in our Refunds, Cancellations & Payments Policy.

21. Liability and Professional Advice

KCO aims to provide AI services with reasonable care and appropriate operational safeguards.

However, AI-generated responses are not guaranteed to be accurate, complete or suitable for every decision.

AI assistance does not replace qualified legal, financial, medical or other regulated professional advice.

Visitors and clients should obtain appropriate independent advice where their circumstances require it.

Specific warranties, remedies and limitations are determined by the applicable agreement and relevant law.

Nothing in this policy excludes liability or rights that cannot lawfully be excluded or restricted.

22. Questions, Complaints and Data Rights

Visitors and clients may contact KCO with questions concerning AI use, inaccurate responses, privacy concerns, suspected misuse or requests for human assistance.

Yumi is KCO’s first point of contact and may assist with the initial handling and routing of such enquiries.

Requests requiring human judgement, authority or legally prescribed handling will be referred for appropriate review.

Individuals may exercise relevant data protection rights as explained in our Privacy Policy.

For contact, email yumi@karelcaslavsky.com.

23. Applicable Law and Policy Updates

This policy is intended to operate in accordance with applicable Czech law and relevant European Union requirements, including applicable data protection and artificial intelligence regulations.

KCO may update the policy as its AI systems, commercial services, security arrangements and legal obligations evolve.

The latest published version will display its revision date.

Existing accepted service agreements are not automatically amended merely because this public policy is revised.

Mandatory legal protections continue to apply where relevant.

CONTACT KCO

Questions about Yumi or our AI services?

Yumi is KCO’s first point of contact for AI enquiries, general policy questions, privacy requests and matters requiring further assistance.

Requests requiring human oversight or approval will be referred through KCO’s appropriate procedures.

Email Yumi