How I use AI in our work together


A plain account of what tools I use, what I use them for, and how your information is handled

AI is part of how I work. I'm not going to pretend otherwise, and I'm not going to hide behind vague language about "advanced tools" or "streamlined workflows." You deserve a straight answer about what's happening when I work on your practice.

Here it is.

Why this distinction matters

When you use AI well — bringing your reflective judgment to evaluate, redirect, and build on what it produces — the interaction between your expertise and AI's processing creates something neither could produce alone. The outcome belongs to the interaction, not to either party.

When you outsource your thinking to AI — accepting its output without engaging your own judgment — there is no real interaction. You get the AI's output. Nothing emerges that wasn't already in the prompt.

The distinction is the difference between a jazz musician using a new instrument and a music library using a search algorithm. One produces something new through interaction. The other retrieves something pre-existing.

The tools I use

I work through Hermes Agent, an AI platform built by Nous Research, running on a paid account. It's one working environment that gives me access to several AI models and a set of tools — web search, research, image generation, and document processing — under a single login.

The practical difference from a single chat window: I can route different kinds of work to different models, and the tools (research, images, analysis) sit alongside the writing rather than in separate apps. It's not magic and it's not secret — it's just a more capable desk.

I use paid accounts across the models and tools I work with. I do not use free-tier tools that train on your inputs, and I do not experiment with new tools on live client accounts.

Images

Images used on this website and in client work come from several sources. Some are from Unsplash. Some are AI-generated using image tools, including Flux, accessed through my working environment. Some are created in Canva through my team member Marynel.

Content Credentials is an emerging industry standard backed by Adobe, Google, Microsoft, Meta, and OpenAI, among others. It embeds provenance metadata directly into image files, recording whether an image was AI-generated and how. Several of the tools I use are implementing this standard. Where an image carries Content Credentials, that information is embedded in the file and can be verified independently at verify.contentauthenticity.org.

I welcome this development — it means I won't need you to take my word for it.

What I use AI for

Drafting and refining your manifesto document. The manifesto is built entirely from what you bring to our discovery sessions — your expertise, your language, your way of engaging with problems. AI helps me synthesise that material into a structured, articulate document. The thinking, the judgment about what's true for your practice, and the quality control are mine throughout.

  • Helping shape and articulate the services you offer, drawing on the same discovery material, so what gets published reflects how you actually work rather than how a template describes you.
  • Drafting initial content that I then read, rewrite, and shape to match your voice as defined in your manifesto.
  • Researching topics and synthesising source material before I write.
  • Interpreting your Plausible Analytics data — the anonymised, cookie-free traffic data that shows how your website is performing.
  • Summarising meeting notes and organising research before strategy sessions.


What I never use AI for

  • Publishing your content without your approval. Every piece goes through my review, any relevant editing, and your sign-off before it goes live.
  • Generating the substance of your manifesto from generic inputs — everything in it comes from you.
  • Replacing experience. The strategic decisions — what to build, what to say, how to position your business — come from decades of working with clients across marketing, not from AI.
  • Sharing client personal details, financial information, client lists, or proprietary business information with any AI tool.

Your information and data protection

I treat your information the way I'd want a partner to treat mine.

I don't enter your name, your clients' names, financial figures, or confidential business details into AI tools. When AI assists with content for your practice, I work with context, not with identifying information.

Your Plausible Analytics data is something I review personally before and after every report. Plausible is a privacy-first platform — no cookies, no personal data, no individual visitor identification. I use AI to help interpret traffic patterns and content performance, working from that anonymised aggregate data, then review the findings myself to identify what's working and where your website, content, or SEO can be improved.

I also choose tools that check their own supply chain. Before my working environment installs a third-party skill or extension, it runs a security scan that flags leaked credentials, personal data, and hidden code before anything runs — one of the reasons I trust it with client work rather than loading tools on faith.

If you ever have a question about what I've shared with an AI tool in relation to your account, ask me. I'll give you a straight answer.

Specialist content

For content that sits outside my own area of expertise — including articles for clients working in medical fertility or engineering and construction — I don't publish without the review and sign-off of a qualified specialist in that field. That may be the client themselves, a specialist editor, or a subject matter expert with relevant credentials. AI assists with drafting and research in these areas; a qualified specialist decides whether it's good enough to publish.

Accountability

Every piece of content that goes out under your name has been read, edited, and approved by me. If an AI draft contains a factual error, a hallucinated claim, or a tone that doesn't match your voice, that's my responsibility to catch — and I do. Recently an AI draft attributed a quote to me that actually came from one of my clients. I recognised the quote, questioned it, and checked it before it went anywhere. That's the job.

AI is a thinking companion I use all day. The thinking, the quality control, and the judgment about what to build and what to leave out, stay mine.

Last updated: 08 September 2026. Questions?  Contact me