How I use AI

AI is part of how I work. I use it to support research, thinking, and delivery across client engagements. This page explains what I use, how I handle client information, and why I’m being open about it.

What I use AI for

I use AI to work more efficiently while maintaining my standards for thinking and rigour.

Research and discovery
AI helps me gather background information, explore industry trends, and synthesise large amounts of material quickly. This means I spend less time on basic research and more time on analysis and strategy.
Drafting and iteration
I use AI to create initial drafts of documents, proposals, and content. I then review, refine, and rewrite. The thinking is mine. AI helps with the mechanics.
Analysis and problem-solving
When working through complex problems, I use AI to explore different approaches, stress-test ideas, or get a second perspective. It’s a useful thinking partner, but the decisions are always mine.
Code and development
For technical work, AI assists with writing code, debugging, and creating documentation. It speeds up routine tasks and helps me maintain quality across projects.
Learning and staying current
Technology moves fast. I use AI to help me understand new tools, platforms, and developments so I can give clients informed, up-to-date advice.

In every case, AI assists the work. It doesn’t replace the thinking or judgment that I bring to every engagement.

What I don’t use AI for

Strategic decisions
AI doesn’t make recommendations or decisions for clients. That’s my job. I use AI to inform my thinking, not to substitute for it.
Final deliverables without review
Client deliverables go through my review process. I check, refine, and approve outputs before delivery.
Sensitive client data
I don’t input personally identifiable information, confidential financials, or proprietary business details into AI tools. More on this below.
Anything I wouldn’t stand behind
Everything I deliver goes through my review process. AI is a tool in that process, not a substitute for professional judgment.

How I handle your data

I take confidentiality seriously. Here’s how I handle client information when using AI tools.

Tools I use

I use the following AI tools, all through commercial API access or local installation:

Claude (Anthropic) and ChatGPT (OpenAI) for reasoning, drafting, and analysis
Claude, OpenRouter, and local models for development
Perplexity for research
Firecrawl for structured web data extraction
Local models that run on my own hardware, where no data leaves my machine

Under their commercial and API terms, Anthropic and OpenAI do not use API input data to train their models. Perplexity maintains a zero data retention policy for its Sonar API and does not use API data for training. OpenRouter does not log prompts by default and offers zero data retention settings, though requests are subject to the underlying model provider’s own data policy. Firecrawl is a web scraping tool that does not store the data it processes beyond what is needed to complete the request.

I use commercial API access specifically because of these protections. I don’t use free-tier AI products for client work, as those tiers may use input data for model training.

For current details on each provider’s data handling, refer to their policies directly:

What I input

When I use AI, it’s typically for general research, structuring ideas, drafting content, or working through technical problems. For example, I might ask AI to help research industry trends, explore approaches to a technical challenge, or debug a piece of code.

I don’t input your company name, confidential details, or information that could identify your business unless it’s already public or destined for publication.

What I don’t do

I don’t use free-tier AI products for client work
I don’t input personally identifiable information, financial data, or confidential business details into AI tools

What I do

I use AI for research, drafting, and analysis involving information that is general, anonymised, or destined for public release
I review all outputs before they reach clients, regardless of what tools were involved
I choose tools based on their data protection policies and commercial terms
I monitor provider policies and update my practices if anything material changes

Prefer no AI?

If you’d rather I didn’t use AI tools on your engagement, let me know. I’m happy to accommodate that where feasible.

If your contract includes specific clauses about AI or data processing, I’m happy to discuss how my practices align. Get in touch.

AI and this website

I use AI when creating content for this website. I set the direction, define the subject matter, and provide the brand and editorial guidelines that shape every piece. AI assists with structuring outlines and producing initial drafts. From there, I refine the structure, reshape the content, and approve the final copy before anything is published.

The subject matter expertise, editorial direction, and quality standard behind this site’s content are mine. AI accelerates parts of the production process.

Human oversight

AI doesn’t work unsupervised in my practice. AI-assisted work goes through my review before it becomes part of client deliverables or published content.

This goes beyond catching errors. It’s about making sure the output reflects my judgment and fits your context. AI can suggest, draft, and assist. Every final output reflects my professional judgment.

Why I’m telling you this

Transparency is a core value in how I work. I’d rather you know how I operate than wonder about it.

AI is a legitimate professional tool when used responsibly. I use it to work more efficiently and deliver more value to clients. Being open about that is part of the trust I want to build with every client.

If you have questions about any of this, I’m happy to discuss it.

How I use AI

Contents

Published Date

Last Modified Date