CLOCOT AI & AUTOMATED SYSTEMS POLICY

Effective Date: October 2, 2026 Last Updated: October 2, 2026

This AI & Automated Systems Policy governs the use of artificial intelligence, automated workflows, Agents, Directors, model providers, synthetic media, and agentic actions within Clocot. It forms part of the agreements governing the Services and should be read together with the Terms of Service, Privacy Policy, and Acceptable Use Policy.

Clocot is designed to let users combine software, data, integrations, AI models, and automation into operational Offices. AI can assist with analysis, drafting, research, planning, content, communications, software tasks, workflow coordination, and authorized actions. AI output can be useful, but it is probabilistic and may be incomplete, inaccurate, outdated, biased, or unsuitable for a particular decision.

1. PURPOSE AND SCOPE

This Policy applies to AI-enabled and automated functionality offered through Clocot, including:

Offices;
Directors;
Agents;
AI Office Architect;
Office Brain and memory;
model routing;
text, image, audio, video, voice, code, search, and data models;
automated workflows;
tool use;
third-party integrations;
autonomous or semi-autonomous actions;
generated and transformed content; and
beta or experimental AI capabilities.

It applies whether the relevant model or tool is provided by Clocot, a third-party provider selected by Clocot, or a provider connected by the user.

2. AI SYSTEM ARCHITECTURE

Clocot may coordinate several components to complete a task. An Office may contain a Director that decomposes work, Agents specialized for particular functions, memory or contextual data, integrations, approval rules, and tools.

An Agent may call another Agent, retrieve information, generate output, invoke a connected service, or return work for human approval.

The specific architecture can change as Clocot develops. Product names describe functional concepts and do not mean that an AI component is a legal person, employee, officer, fiduciary, or licensed professional.

3. MANUAL, HYBRID, AND AUTOPILOT MODES

Clocot may support different operating modes.

Manual Mode requires users to initiate or approve relevant actions.

Hybrid Mode allows some steps to execute automatically while defined checkpoints require human approval.

Autopilot Mode may allow authorized workflows to continue without case-by-case approval, subject to configured permissions, limits, technical safeguards, and available controls.

The user remains responsible for selecting a level of automation appropriate to the task and risk.

4. HUMAN CONTROL AND OVERSIGHT

Automation does not eliminate the need for judgment.

Users should apply meaningful human review when output or actions could materially affect:

finances;
legal rights;
health or safety;
employment;
access to services;
contracts;
public communications;
reputation;
regulated activity; or
other significant interests.

Clocot may require human approval for particular functions, may prevent an automated action, or may introduce additional verification when reasonably necessary for safety, security, compliance, or product integrity.

5. AUTHORIZATION FOR AUTOMATED ACTIONS

An Agent or Director may act only within the technical permissions available to the Account, Organization, Office, workflow, integration, or tool.

By enabling an automated workflow, you authorize Clocot to execute the categories of actions that the workflow is configured to perform.

You are responsible for reviewing:

connected accounts;
tool permissions;
recipients;
spending or usage limits;
approval points;
schedules;
triggers;
data access;
publishing permissions; and
escalation rules.

Revoking a permission may stop future actions but may not reverse an action already transmitted to an external service.

6. THIRD-PARTY AI PROVIDERS

Clocot may route tasks to third-party AI Providers.

Different providers may have different:

capabilities;
safety systems;
availability;
retention practices;
service regions;
accuracy characteristics;
latency;
pricing;
intellectual-property terms; and
data-use rules.

Where a user connects a personal or business API key, that provider may process data under the user’s direct agreement with that provider.

Clocot may replace or reroute providers when necessary for reliability, quality, availability, cost, safety, or product development.

7. PROMPTS, CONTEXT, AND CUSTOMER DATA

AI features may process prompts, uploaded files, memory, connected-account data, tool results, Customer Content, and contextual information reasonably necessary to perform the requested task.

Users should not submit information they are not authorized to process.

Sensitive information should be used only where necessary, lawful, and appropriate for the relevant feature.

The Privacy Policy explains how LexLupi processes personal information and how third-party providers may be involved.

8. GENERATED OUTPUT

AI-generated output may include text, images, video, audio, voice, code, structured data, recommendations, summaries, classifications, plans, drafts, or other results.

Generated Output may:

contain factual errors;
omit relevant information;
repeat patterns from training data;
be similar to output generated for another user;
contain insecure code;
produce broken links or fabricated citations;
be unsuitable for publication;
require fact-checking;
require rights clearance; or
fail to reflect the latest information.

Users are responsible for deciding whether output is appropriate for their intended use.

9. HALLUCINATIONS AND ACCURACY

Generative AI may confidently produce incorrect information.

Clocot does not guarantee that an AI response is accurate, complete, current, unique, lawful, or fit for a particular purpose.

Where accuracy matters, users should verify the result against authoritative sources, source documents, qualified professionals, or other reliable evidence.

The presence of citations, links, calculations, or structured reasoning does not by itself guarantee correctness.

10. PROFESSIONAL AND HIGH-IMPACT USE

Clocot is not a substitute for a licensed professional.

AI output should not be treated as final medical, legal, financial, tax, investment, accounting, engineering, or other regulated professional advice merely because the system can generate information in those fields.

A qualified human should review material professional or high-impact decisions.

Customers are responsible for determining whether their use requires licensing, notices, recordkeeping, impact assessments, human review, explainability, or other regulatory controls.

11. AUTOMATED DECISIONS ABOUT PEOPLE

Clocot must not be used to make unlawful or unfair decisions about individuals.

Users must not rely solely on automated processing for decisions that produce legal or similarly significant effects where applicable law requires human involvement, notice, consent, explanation, contest rights, or other safeguards.

Examples can include employment, housing, lending, insurance, education, healthcare, essential services, and other regulated contexts.

12. FAIRNESS AND NON-DISCRIMINATION

Users must not configure Clocot to discriminate unlawfully against individuals or groups.

Where AI is used to rank, classify, prioritize, or recommend people, the user should consider:

data quality;
representativeness;
protected characteristics;
proxy variables;
historical bias;
false positives;
false negatives;
accessibility; and
the consequences of error.

13. SYNTHETIC CONTENT

Clocot may generate or transform synthetic text, images, audio, video, voices, avatars, or other media.

Users are responsible for ensuring that publication or distribution complies with applicable laws, platform rules, publicity rights, intellectual-property rights, privacy rights, advertising rules, and disclosure obligations.

Where applicable law or the receiving platform requires disclosure that content was AI-generated or materially manipulated, the user must provide that disclosure.

14. IMPERSONATION, VOICE, AND DEEPFAKES

Users must not use Clocot to deceptively impersonate another person, fabricate authorization, create fraudulent identity evidence, or mislead people about the source of a communication.

Synthetic voice, face, or likeness tools may be used only where the user has appropriate rights, permissions, or another lawful basis.

Clocot may restrict features involving identity replication, biometric identification, or deceptive synthetic media.

15. AI-GENERATED COMMUNICATIONS

When AI drafts or sends emails, messages, posts, calls, or other communications, the user remains responsible for:

recipient consent where required;
sender identification;
required disclosures;
unsubscribe or opt-out mechanisms;
recording consent;
telemarketing rules;
anti-spam laws;
platform rules; and
the substance of the communication.

Automated generation does not convert an unlawful communication into a lawful one.

16. CONTENT AND RIGHTS REVIEW

AI output can unintentionally resemble third-party material.

Before commercial publication or distribution, users should consider whether output requires review for:

copyright;
trademark;
publicity rights;
confidentiality;
privacy;
licensing;
defamation;
advertising claims; and
contractual restrictions.

As between the user and LexLupi, rights in inputs and outputs are governed by the Terms of Service, subject to applicable law and third-party provider terms.

17. MODEL TRAINING AND PRODUCT IMPROVEMENT

LexLupi’s handling of Customer Content for service improvement and model-related purposes is described in the Privacy Policy.

Clocot does not treat Customer Content as unrestricted training data.

Third-party AI Providers may have separate retention or improvement practices. Those practices can depend on whether Clocot manages the provider relationship or the customer uses its own provider account.

18. MEMORY AND PERSONALIZATION

Office Brain and related features may preserve context, preferences, historical information, and learned workflow patterns.

Persistent memory can increase consistency but can also preserve outdated or incorrect information.

Users should periodically review important persistent instructions and should correct or delete information that is no longer appropriate.

19. SAFETY CONTROLS

Clocot may use technical and operational controls designed to reduce misuse, including:

permission limits;
rate limits;
model safety systems;
approval checkpoints;
content filtering;
monitoring;
abuse detection;
credential controls;
restricted tools;
blocked actions; and
account suspension.

No safety system is perfect. Users remain responsible for lawful and appropriate use.

20. PROHIBITED AI USES

AI features may not be used for conduct prohibited by the Terms of Service or Acceptable Use Policy.

This includes unlawful activity, fraud, phishing, malware, deceptive impersonation, unlawful surveillance, exploitation, unauthorized credential access, unlawful discrimination, and other prohibited uses.

We may impose additional restrictions on features that present elevated legal, safety, or abuse risk.

21. BETA AND EXPERIMENTAL AI

Beta, preview, Lab, and experimental AI features may:

change without notice;
produce unstable results;
have additional limitations;
use temporary providers;
have reduced availability; or
be discontinued.

They should not be used as the sole control for mission-critical or safety-critical processes unless expressly approved for that purpose.

22. MONITORING AND ENFORCEMENT

We may investigate suspected misuse and may limit, suspend, or terminate access where reasonably necessary to protect users, third parties, Clocot, providers, or the integrity of the Services.

Where legally permitted, we may preserve relevant records for security, abuse investigation, legal compliance, or dispute resolution.

23. CHANGES TO THIS POLICY

AI technology and regulation evolve quickly.

We may update this Policy to reflect new models, features, safety controls, legal obligations, or provider practices.

Material changes will be communicated where required by law or contract.

24. CONTACT

Questions about this Policy may be sent to office@clocot.com.

LexLupi LLC Privacy and Legal Contact: Tatjana Sindjelic U.S. registered office: 131 Continental Dr, Suite 305, Newark, DE 19713, United States Serbian principal business address: Save Kovacevic 29/4, 34000 Kragujevac, Serbia