AI Technology Terms

v1. Last updated 18 June 2026

These Velixo AI Technology Terms (the “AI Terms”) govern the access to and use of cloud-enabled artificial intelligence features, connectors, applications, plug-ins, and capabilities made available by Velixo (collectively, “Velixo Intelligence” or “VI”).

These AI Terms form a binding legal agreement between Velixo and the entity or individual accessing VI (“User”) and are incorporated by reference into the Velixo End User License Agreement (“EULA”). In the event of any conflict between the EULA and these AI Terms, these AI Terms shall control strictly with respect to User’s use of Velixo Intelligence.

1.  DEFINITIONS

• “AI Input” means any financial data, Enterprise Resource Planning (ERP) metrics, prompts, natural language instructions, queries, text, or database schemas submitted or routed by User (or Authorized Users) into Velixo Intelligence.
• “AI Output” means the text, numbers, formulas, financial summaries, reports, insights, visualizations or automated data structures generated and returned to User by Velixo Intelligence.
• “ERP Write-Back” means the automated insertion, modification, deletion, or synchronization of data directly within a User’s connected Enterprise Resource Planning (ERP) system, triggered or orchestrated by Velixo Intelligence based on an AI Input.
• “Model Provider” means third-party artificial intelligence infrastructure and large language model platforms utilized by the Velixo AI Gateway network to process AI Inputs and generate AI Outputs, as identified in the AI Services Schedule.
• “Velixo AI Services Ecosystem” means the collective network of multi-directional interfaces managed by Velixo, including local Excel add-ins, standalone web or desktop applications, and custom connectors or plugins deployed within third-party AI marketplaces (such as OpenAI Custom GPTs, Anthropic Claude Artifacts/Apps, or native marketplace ecosystems).
• “Velixo AI Gateway” means the cloud-hosted infrastructure layer operated by Velixo that acts as a secure, intermediate proxy between the User’s environment (including the Software or any external marketplace application) and the third-party Model Providers. The Velixo AI Gateway orchestrates the secure routing, ingestion, system-prompt augmentation, and proprietary transient diagnostic logging of AI Inputs and AI Outputs.
• “AI Services Schedule” means the schedule maintained by Velixo that identifies, and that Velixo may update from time to time, the current AI Services, AI Features, Model Providers, subprocessors, and applicable data retention periods for Velixo Intelligence and any other proprietary AI-based product.

 2. ARCHITECTURE & MULTI-LAYER DATA GOVERNANCE

User acknowledges and agrees that Velixo Intelligence does not process data locally. All data passing through the Velixo AI Services Ecosystem is governed by a secure, multi-layered data flow designed to guarantee transient, controlled data handling:

(a) Velixo AI Gateway layer. The Velixo AI Gateway, which Velixo owns and operates, routes AI Inputs and AI Outputs to the Model Provider and retains a transient diagnostic log of AI Inputs and AI Outputs for the limited rolling period stated in the AI Services Schedule, after which the log is automatically purged. Access to these logs is restricted to a limited number of Velixo engineers and technical leads for troubleshooting and for improving the Velixo Intelligence agents. Velixo does not use AI Inputs or AI Outputs to train or fine-tune any large language model.

(b) Model Provider layer. AI Inputs are processed by one or more third-party Model Providers identified in the AI Services Schedule. Each Model Provider is contractually committed not to use AI Inputs or AI Outputs to train its models. A Model Provider may retain AI Inputs and AI Outputs for a limited period for abuse monitoring as described in the AI Services Schedule. User acknowledges that this Model Provider retention is required by the Model Provider, applies regardless of any Velixo-level setting, and cannot be opted out of.

(c) Sensitive data. User acknowledges that AI Inputs and AI Outputs may include financial and other sensitive data, and that such data is logged transiently at the layers and for the periods described above and in the AI Services Schedule.

3. THE ENTERPRISE TRUST & ZERO-TRAINING COVENANT

• (a) Strict Model Training Prohibition. Velixo does not use AI Inputs or AI Outputs to train, retrain, or fine-tune any large language model. Each Model Provider is contractually committed not to use AI Inputs or AI Outputs to train its foundational models, as described in the AI Services Schedule.
• (b) Platform Isolation. User data is not comingled with other customers’ data and is not used to train or improve any foundational model. Software updates to Velixo Intelligence are achieved exclusively by modifying global infrastructure configurations, underlying processing algorithms, and programmatic system prompts to improve general-purpose problem-solving capability without absorbing user context.

4. DATA LOGGING CONTROLS AND SUPPORT LIMITATIONS

Opting out of Velixo Intelligence is available any time , at the user-level or at the organization-level through organization-level administrative controls. From general availability of Velixo Intelligence. Opting out of log storage does not affect the Model Provider abuse-monitoring retention described in Section 2 and the AI Services Schedule, which the User cannot opt out of.

In the event User exercises this log opt-out preference, User explicitly acknowledges that Velixo’s engineering and support teams will lose the technical capacity to investigate, reproduce, audit, or resolve software bugs, agent malformations, or performance issues occurring within Velixo Intelligence. Velixo disclaims all technical support obligations, Service Level Agreement (SLA) metrics, and warranty remedies for any service failures directly or indirectly arising from the unavailability of diagnostic logs due to a User opt-out.

5. ERP WRITE-BACK OPERATIONAL GOVERNANCE & SAFEGUARDS

ERP Write-Back is disabled by default and is enabled only through a separate, explicit administrative opt-in by the User. If User chooses to authorize Velixo Intelligence to execute ERP Write-Back functions, User explicitly accepts full operational liability and agrees to the following structural terms:

• (a) Dual-Authorization Principle. Velixo Intelligence is configured to require clear confirmation, human intervention, or programmatic administrative approval safeguards before pushing any structural, transactional, or financial records into a connected ERP database. User agrees never to bypass, disable, or programmatically circumvent these approval steps.
• (b) Ultimate Human Accountability. Velixo Intelligence operates on probabilistic machine learning models that interpret natural language intents. User acknowledges that natural language instructions can result in unintended code interpretation or data mapping.
• (c) Disclaimer of Write-Back Liability. User retains sole and exclusive responsibility for reviewing and confirming the validity, formatting, precision, and systemic impact of any ERP Write-Back action prior to granting final approval. Velixo disclaims all liability for corrupt database states, erroneous ledger postings, automated overwrites, structural system failures, or business disruptions arising out of an ERP Write-Back action executed under the systemic instruction of an AI Input.

6. THIRD-PARTY AI PLATFORMS, CONNECTORS, AND MARKETPLACES

When User accesses Velixo Intelligence through external environments – including third-party AI platforms or marketplaces identified in the AI Services Schedule – the following terms apply:

• (a) Scope of Velixo Liability. Velixo’s security, privacy, and data governance guarantees apply strictly to data moving within the Velixo AI Gateway network.
• (b) Boundary of Third-Party Terms. The moment an AI Output is pushed or rendered back onto a third-party platform’s own interface, that data becomes subject to the User’s independent commercial data privacy agreement with that specific platform vendor. User is solely responsible for vetting and securing its own regulatory compliance within third-party AI interfaces.

7. LIMITATION OF LIABILITY AND AI OUTPUT DISCLAIMER

The Disclaimer of Warranty and the Limitation of Liability set out in the EULA, including the liability cap stated in the EULA, apply in full to Velixo Intelligence, AI Inputs, and AI Outputs. Nothing in these AI Terms increases Velixo’s aggregate liability beyond the cap stated in the EULA.

• (a) Probabilistic Nature of Outputs. User acknowledges that AI Outputs are generated using statistical patterns and mathematical probabilities. Velixo makes no representation, warranty, or guarantee regarding the accuracy, absolute mathematical correctness, compliance, or completeness of any AI Output.
• (b) Professional Oversight. AI Outputs are provided for operational speed and automation convenience only and do not constitute certified financial analysis, auditing, tax advice, or regulatory compliance verification. User must subject all AI Outputs to human professional review before leveraging them for legal or corporate decision-making.
• (c) Exclusion of Consequential Damages. VELIXO SHALL NOT BE LIABLE UNDER ANY CIRCUMSTANCES FOR CORPORATE LOSS OF PROFITS, ERRONEOUS TAX FILINGS, REPORTING DISCREPANCIES SUBMITTED TO CAPITAL MARKETS, REGULATORY FINES, OR FINANCIAL AUDIT FAILURES DISCOVERED WITHIN THE USER’S GENERAL LEDGER SYSTEM RESULTING FROM THE USE OF VELIXO INTELLIGENCE.