Essential Content AI policy

Version	1.0	Owner	Sean O’Meara, Managing Director
Approved	14 October 2026	Next review	14 January 2027
Applies to	All employees, freelancers, contractors and partners working on behalf of Essential Content Ltd (company number 09737309)

1. Summary

This document outlines our organisational policy on AI use.

We’ve written this document to provide context and detail about how, why and when Essential Content uses AI.

All work is created by humans, edited by humans and approved by humans.

We are alert to both the potential and risks of AI use and we hope this document sets out clearly how we balance the opportunities with our obligations to treat clients well.

The majority of our work is created to be read and used by humans.

Some of our work is created to improve search engine performance and large language model (LLM) retrievability. This work is always created by humans but is optimised to make it easier for search engines and large language models to retrieve, understand and index.

We never use AI for:

• Writing
• Editing
• Optimising
• Fact-checking
• Sourcing

2. Purpose

This policy explains how Essential Content uses, and does not use, artificial intelligence (AI) tools. It exists to:

• Protect the quality, accuracy and originality of the work we deliver
• Protect client confidentiality and personal data
• Be open with clients about how their work is produced
• Explain how AI tools can help us produce elite content at scale

3. Scope

This policy covers generative AI tools (for example ChatGPT, Claude and Microsoft Copilot), AI features built into other software (for example grammar checkers, transcription, search, design tools) and any tool that generates text, images, audio or video.

It applies to client work, new business processes, internal documents and marketing material.

4. Definitions

• Generative AI: Paid and free tools powered by machine learning. This includes standalone tools and agents, like Claude and ChatGPT as well as AI integrations in other software.
• Client material: Any content, whether for publication or not, supplied to us by a client or created by us for a client.
• Personal data: as defined by UK GDPR.
• Human-created work: Content for publication, emails, strategy documents and any form of content we provide to a client whether defined by a scoping document or ad-hoc will always be attributable to a named person working for Essential Content.

5. Our principles

1. Human authorship. Every piece of client work is researched, written and edited by a named person or named people on our team.
2. Human judgement. People make every editorial, strategic and compliance decision.
3. Confidentiality. Client material never goes into a tool that could store it or train on it without written permission.
4. Accuracy. Nothing is published or handed over without a person checking every fact, quote and figure against a primary source.
5. Transparency. We tell clients how their work is produced and answer honestly if asked.
6. Accountability. Where requested, we’re able to show work logs, version control and internal audit trails showing the lifecycle of a document, enabling our clients to examine how we produced a piece of work.

6. What we use AI for

Permitted use	Conditions	Approved tools
Transcribing interviews and meetings	With active consent of participants or prior written consent	Fireflies.AI Google Gemini
Spelling and grammar checks	Suggestions reviewed, not auto-accepted	Google Gemini Claude
Background research and finding sources	Every fact verified against an original source	Claude ChatGPT Google Gemini
Internal admin, scheduling, summarising our own notes	No client material, unless the client has given permission	Google Gemini Claude Monday.com
Project documentation, change-logs and version control	Client material accessible with permission	Google Gemini Claude Monday.com

7. What we don’t use AI for

• Writing or editing copy, articles, press releases, thought leadership or other public facing content.
• Generating quotes and insights
• Creating images or audio presented as real
• Making compliance or regulatory judgements
• Pitching to or corresponding with journalists, editors, producers or other media colleagues

8. Client material and personal data

• Client material and personal data must not be entered into any AI tool unless it is on the approved list in section 6 and the client has agreed in writing.
• Approved tools must have training on our data switched off and meet UK GDPR requirements.
• Confidential, unpublished or market-sensitive information (for example results, deals, crisis matters) must never be entered into any AI tool.
• Where possible, we will use our best endeavours to anonymise any client data before entering it into an AI tool.
• See our privacy policy at essentialcontent.co.uk/privacy-policy/.

9. Client choice and consent

Clients always have control over what we do with their data, content and information.

If a client wants their content to never come into contact with an AI tool, regardless of AI model training settings, we have processes and procedures in place to make sure this happens. For example, we have separate digital working environments where all AI tools are either disabled or switched off.

All team members work in a shared collaboration environment and where required we are able to prevent materials being copied or exported from that environment.

10. Freelancers and external partners

All suppliers to Essential Content are subject to a supplier code of conduct which outlines their obligations as they relate to the use of AI. We also have internal tooling and processes to protect content from being exposed to tools outside of our approved tool list.

11. Accountability

• Policy owner: Sean O’Meara, managing director
• Questions or concerns: sean.omeara@essentialcontent.co.uk

12. Approved tools register

Tool	What it’s used for	Data/training setting	Approved by	Date
Google Gemini	Video and audio call transcription, research, grammar and spellchecking, document retrieval and summary, currency conversion and document formatting.	Paid plan	Sean O’Meara	14 October 2026
Claude	Version control generation Changelog generation Internal admin Compliance flagging	Max Plan, Claude is not permitted to use any content for training	Sean O’Meara	14 October 2026
Monday.com custom agents	No current uses	Paid plan	Sean O’Meara	14 October 2026
Microsoft Copilot	No current uses	Paid plan	Sean O’Meara	14 October 2026
Fireflies.AI	Transcribing interviews and meetings	Paid plan	Sean O’Meara	14 October 2026

13. Review

We review this policy every three months or sooner if tools, regulation or client needs change. Last reviewed: 14 October 2026.

Frequently asked questions

Do you use AI to write content?

No. We have a talented team of writers, editors and strategists who we rely on to produce quality content. We have yet to find a way to use AI that creates better content than our team.

So do you use AI at all?

We use AI to improve our processes and for support with automatable tasks in the review and documentation process. For example, we have custom agents that can tell us whether a piece of content requires a certain disclaimer, which we can retrieve from a human-approved catalogue of disclaimers.

We also have AI agents trained to flag compliance risks, such as undue prominence of product benefits or unfair comparisons. Every flag is verified by a human.

What does “created by humans, for humans” actually mean?

AI supports our backend processes, humans do the creative work. In practice, this means that any work produced for a client or for our own channels is written, edited, checked and approved by a human.

Will my information be put into ChatGPT or another AI tool?

Not without prior written permission. All of the AI tools we use are set to not allow AI to use the material we work on for training.

Can I ask you to use AI on my project?

You can always ask and we will always consider the specific merits of AI use in a project. On the whole, we will decline any request to use AI for writing.

Can I ask you not to use AI at all?

Yes. And we will gladly comply, with proof of compliance. In practice, this means that we may not be able to deliver some of the operational efficiencies that AI can deliver, such as those described above.

How do you check facts and figures?

Our fact-checking systems were developed by the award-winning journalists in our team. We use double-sourcing and have an internally approved list of primary sources.

Do your freelancers follow the same rules?

Yes. Everyone in our team is subject to the same rules governing our use of AI. In the vast majority of cases, our systems prevent non-compliance.

How do you handle regulated content, such as financial promotions?

AI supports some internal compliance processes relating to regulated content. For example, we have custom agents that can flag risks and non-compliance in a piece of content. We also have custom agents that can identify unmet regulatory requirements in content.

Do you use AI to pitch journalists?

Never.

Do you create content for AI search (AEO)?

Yes. Many of our clients ask us to help them create content that serves AI / LLM retrievability as well as traditional search engines.

Why is there a page on your website for AI assistants?

This page is designed and created for AI assistants, so they can access the most up-to-date and accurate information about Essential Content.

Who do I contact with a question about this policy?

Sean O’Meara – sean.omeara@essentialcontent.co.uk