Ai Policy

Responsible AI Use Policy

Updated: May 2026 | Policy owner: Head of Operations

1. Introduction

CommsKit uses artificial intelligence (AI) as a managed support tool to improve efficiency, consistency, quality control, and creative development in our work.

AI supports our team – it does not replace human judgement, professional responsibility, or client approval processes. All AI-assisted work is reviewed, edited, and approved by a CommsKit team member before it is shared externally.

This policy explains how CommsKit uses AI in client work, how we protect client information, and how we respond to clients with specific confidentiality, regulatory, or data protection requirements.

2. Our AI principles

CommsKit’s use of AI is guided by the following principles:

Human accountability: AI may assist our process, but humans remain responsible for decisions, quality, accuracy, and final delivery.

Confidentiality: We do not input confidential, sensitive, or regulated client information into external generative AI tools without written client permission and a risk assessment approved by the Head of Operations.

Transparency: We disclose our use of AI tools where relevant and work with clients who have specific AI restrictions or disclosure requirements.

Data minimisation: We only use the information required to complete the task and avoid unnecessary sharing of client material with AI tools.

Quality control: AI outputs must be checked for accuracy, tone, relevance, bias, copyright risk, and factual reliability before use.

Appropriate use: AI must be used to support professional delivery, not to bypass expertise, due diligence, review, or approval.

3. Why we use AI

CommsKit may use AI tools to:

  • improve proofreading, grammar, clarity, and consistency;
  • check brand tone and messaging alignment;
  • support meeting notes, summaries, and action tracking;
  • assist with research synthesis and briefing summaries;
  • support copy development, rewriting, structuring, and editing;
  • assist with campaign, content, and design ideation;
  • support layout, storyboard, visual concept, and production planning;
  • improve workflow management and internal quality checks;
  • help structure content for presentations, e-learning, digital platforms, websites, and marketing material;
  • automate repetitive administrative or production tasks where appropriate.

AI is used as part of the process, not as the final unchecked output.

4. How AI may be used in client work

Examples of AI-assisted work may include:

  • content outlines, rewrites, and copy refinement;
  • proofreading and quality assurance;
  • tone-of-voice checks against approved brand or communication guidance;
  • summarising long-form content into usable working notes;
  • identifying key messages from approved reference material;
  • concept development, moodboard support, and visual exploration;
  • presentation, brochure, emailer, and campaign structure support;
  • UI/UX content planning and wireframe support;
  • e-learning content structuring, scripting, and learning-flow support;
  • meeting transcription, summaries, and action lists;
  • internal workflow and project management support.

AI-generated or AI-assisted material is not sent to a client without human review and approval by a CommsKit team member.

5. Tools we may use

CommsKit uses a range of paid, business, professional, and embedded AI-enabled tools as part of our workflow. These may include, but are not limited to:

  • generative AI and large language model tools;
  • AI-supported meeting, transcription, and note-taking tools;
  • Microsoft 365 and Microsoft Copilot tools;
  • proofreading, editing, and writing-support tools;
  • design, image, video, layout, and production tools with embedded AI features;
  • workflow automation and productivity tools.

The exact toolset may change as technology, security standards, and client requirements evolve. Where available, CommsKit uses business, team, or professional versions of AI tools and applies the privacy and model-training controls available within those tools.

CommsKit does not permit staff to use personal, unmanaged AI accounts for client work where a CommsKit-approved system is required.

We describe our AI use by tool category rather than a live tool list, as our toolset evolves. Clients who require additional information – including data-processing detail, sub-processor identification, or account-specific restrictions – should contact our Head of Operations directly. We will provide relevant information and agree handling steps on a case-by-case basis.

6. Privacy, confidentiality, and data protection

CommsKit’s staff and approved subcontractors are required to protect client information and follow confidentiality obligations.

We do not input any of the following into external generative AI tools without written client permission and a risk assessment approved by the Head of Operations:

  • confidential client information;
  • personal data;
  • customer data;
  • employee data;
  • financial records;
  • regulated information;
  • unpublished sensitive reports;
  • internal systems data;
  • proprietary datasets;
  • commercially sensitive or access-controlled material.

Where AI tools are used, the material shared is limited to what is necessary and appropriate for the task. This may include public-facing material, approved project content, brand guidance, tone-of-voice examples, published marketing material, or non-sensitive working content.

When working with clients in regulated sectors, or clients who provide specific AI usage requirements, CommsKit will apply account-specific handling instructions as agreed.

7. Client-specific AI restrictions and disclosure requirements

Some clients have additional requirements relating to AI use, confidentiality, data residency, sub-processors, automated decision-making, meeting recording, or disclosure obligations.

Clients should provide any AI-related restrictions, contractual requirements, or approval processes at the start of a project, or as soon as they become aware of them.

Where a client notifies CommsKit of specific restrictions, we will confirm the agreed handling steps in writing. These may include:

  • restricting the types of AI tools used on the account;
  • limiting AI use to specific support functions;
  • excluding certain categories of client material from AI-assisted workflows;
  • providing additional disclosure about AI-enabled tools;
  • adding explicit meeting-by-meeting notification where AI-supported note-taking is used;
  • requiring additional internal review before delivery;
  • agreeing alternative manual processes where feasible.

If restrictions are introduced after a project has been scoped or started, they may affect the timeline, workflow, cost, or delivery approach. CommsKit will discuss any such impact with the client before proceeding.

8. Meeting recording, transcription, and AI note-taking

CommsKit may use meeting recording, transcription, or AI-supported note-taking tools to help capture accurate summaries, decisions, and actions.

Where these tools are used, CommsKit will notify participants before or at the start of the meeting. Notification may be given through meeting platform indicators, email disclosure, verbal confirmation, or other appropriate notice.

For clients with stricter requirements, explicit verbal confirmation is given at the start of every meeting where AI-supported note-taking or transcription is active.

Meeting notes and summaries are working records and remain subject to CommsKit’s confidentiality and data handling obligations.

9. Internal AI assistants and configured workspaces

CommsKit may use internal AI assistants, custom GPTs, or similar configured workspaces to support specific workflows, accounts, or production tasks.

These tools are used internally and are not distributed to clients or the public unless specifically agreed.

Client-related internal assistants are configured using approved reference material only – such as public-facing content, approved marketing copy, brand guidance, tone examples, project instructions, or internal CommsKit guidance. They are not configured with confidential, personal, regulated, or sensitive client material unless the client has provided written permission and the use has been reviewed by the Head of Operations.

Access to internal assistants is limited to authorised CommsKit employees and approved subcontractors who are bound by confidentiality obligations.

10. Intellectual property and AI-generated assets

CommsKit treats AI-generated or AI-assisted material with care in respect of intellectual property, copyright, and rights ownership.

Where AI-generated images, media, copy, or other assets are included in a deliverable, CommsKit will:

  • confirm the material is suitable for the intended use;
  • carry out any additional human editing or original creation required;
  • assess whether licensing, copyright, or usage restrictions apply before use;
  • flag to the client where AI-generated assets have been used in a deliverable, where relevant and where the client has not otherwise indicated a preference.

If originality, exclusivity, or full rights ownership is required, clients should raise this at briefing stage so that CommsKit can recommend an appropriate production route.

11. Accuracy, review, and approval

AI tools can produce inaccurate, incomplete, biased, or misleading outputs.

For this reason, all AI-assisted content is reviewed by a CommsKit team member before it is used or shared. Depending on the task, review covers:

  • factual accuracy;
  • brand tone and messaging;
  • spelling, grammar, and formatting;
  • visual suitability;
  • source reliability;
  • legal, regulatory, or compliance sensitivity;
  • copyright or licensing risk;
  • alignment with the client brief.

AI is not a substitute for client approval, technical review, legal review, compliance review, or subject-matter expert input where those are required.

12. Responsible AI governance

CommsKit is committed to responsible AI adoption. We actively train our team on appropriate AI use, confidentiality boundaries, quality control, and risk awareness.

Our AI governance continues to evolve as tools, legislation, client requirements, and best practice develop.

CommsKit operates in line with applicable data protection legislation, including South Africa’s Protection of Personal Information Act (POPIA) and the UK General Data Protection Regulation (UK GDPR) and EU General Data Protection Regulation (EU GDPR) where EU or UK personal data is involved. This policy is reviewed at least annually and whenever significant changes occur in our toolset, client base, or the regulatory environment.

CommsKit is also developing more controlled internal AI systems and workflows, supported by technical oversight, to improve consistency, governance, and data handling across our business.

13. Questions or account-specific requirements

Clients with questions about this policy, or who require specific AI restrictions for their account, should contact CommsKit’s Head of Operations through our contact form

CommsKit will review reasonable client-specific requirements and agree practical steps – confirmed in writing – that protect client information while allowing us to deliver work efficiently and professionally.

CommsKit – Responsible AI Use Policy – May 2026 – Confidential