AI Policy

TEG Engineering AI Usage, Coding Creation & IP Integrity Standard Operating Procedure (“SOP”)

Our philosophy:

  • We encourage the use of TEG-approved AI tools to accelerate our roadmap, reduce toil, and enhance creativity. All AI-assisted coding must be human-led and will be documented to log human authorship. This framework ensures we move fast while protecting the Intellectual Property (IP) that makes TEG unique.

1. Purpose and Scope

This SOP applies to:

  • All technology staff.
  • All third-party contractors and external developers with access to TEG systems.

2. Approved AI Tooling

2.1 Approval Requirement

No new AI coding tool may be used unless explicitly approved in accordance with the vendor review detailed below.

2.2 Vendor Review

All AI coding assistants must be approved for use by these TEG teams:

  • Technology (Architecture)
  • Security (CISO)
  • Legal (Legal Tech review comprising contract and privacy/data compliance review).

2.3 Authorisation Criteria

We will only use Enterprise-tier tools that:

  • Contractually guarantee Zero Data Retention
  • Confirm that TEG code is not used to train external models
  • Provide IP indemnification
  • Meet TEG security requirements.

Use of personal, consumer-tier, or non-approved AI tools to assist with TEG code is prohibited.

4. Code Segmentation and AI Tool Usage

 

Category

Definition

Usage Standards

Tier 1: IP Critical

Core, proprietary codebases, services, or algorithms that provide TEG with a direct competitive advantage or handle highly sensitive business logic.

Examples include:

• Ticketing allocation and seating algorithms

• Dynamic pricing engines

 

Audit logs of user prompts, with their suggested code, and whether or not this code was accepted and/or amended, to be maintained and stored within Gemini Code Assist’s native “Cloud Monitoring” capabilities (details here).

Evidence of Human Authorship:

To legally protect TEG’s IP, it is recommended developers leave an audit trail proving they defined the structural logic (“the blueprint”) prior to using AI. This should be documented via:

• Jira/Confluence: Technical specifications and architectural decisions should be documented before prompting AI for large code blocks.

• Pull Requests: PR descriptions should explicitly state the human-directed logic and note where AI was used for heavy generation.

Do not blind-accept large blocks of code. You must review, refine, and adapt the code to TEG’s specific standards. This human refinement is legally required to make the code defensible as TEG property.

Your professional judgment is what secures our IP. When you review and change a suggestion, refine a variable name to match our standards, or tweak logic to improve performance, you are providing authorship that makes that code uniquely ours. These refinements are what makes our tech defensible and valuable.

Tier 2: General

Standardised code, boilerplate, and supporting services that follow industry-standard patterns rather than proprietary TEG innovation (e.g., UI components, integrations, unit tests, and non-core services)

No tier-specific guidelines. AI tools to be used in accordance with this broader policy.

5. Awareness & Attestation

Before using AI coding tools, each developer must read and acknowledge this SOP. If anything is unclear, or you are in doubt, please speak with your Manager or project lead.

6. Review & Updates

This SOP will be reviewed periodically by Technology, Security and Legal to reflect:

  • Changes in AI technology and approved tools
  • Developments in copyright law
  • Security risk evolution
  • TEG’s platform growth.
Scroll to Top