How I use Artificial Intelligence in my work

Artificial Intelligence (AI) offers genuine opportunities to enhance the quality and efficiency of my work, from handling routine tasks through to supporting more complex analysis and synthesis. These opportunities come with responsibilities. My aim is to use AI in ways that amplify my capabilities while upholding the professional standards, human expertise, and care that are at the heart of what I do.

While AI supports my work, it does not replace my professional judgement, expertise, or the personalised attention I bring to every client relationship.

Above all, I believe in being transparent with the people who use my services about the tools and methods I use, including those driven by AI.

Which AI tools I use?

I use a number of AI powered tools. These are listed below in order of the frequency of my use at time of writing:

• Claude from Anthropic which I use for a range of support tasks (regular basis: described in more detail below).

• Fathom for recording and transcribing meetings and interviews (on occasion and only with consent).

• AI tools built into the Office 365 package (occasional use for transcription support within research and evaluation services).

How I use Claude from Anthropic

I use Claude from Anthropic (pro plan) to:

• Assist with structuring written documents and, less commonly, producing initial drafts.

• Support research to help me stay current in my field.

• Organise complex information into clear, accessible formats and summaries.

• Review and order personal notes to inform and guide my practice.

• Improve administrative and business efficiency, including developing, reviewing, and formatting documents.

AI in evaluation and research work

When using AI in evaluation and research-related work, I apply additional care and a higher standard of scrutiny, reflecting the responsibilities I hold to clients, commissioners, and the people whose experiences evaluation seeks to understand.

I align my practice with the UK Evaluation Society's AI in Evaluation: Good Practice Guidelines for Practitioners (November 2025). These set out four core principles that guide my approach: transparency and accountability; human control and proportionate use; active risk management and harm prevention; and quality assurance and verification. In practice this means I:

• Disclose AI use clearly in evaluation reports and proposals, including what tools were (or plan to be) used and how outputs were verified.

• Retain full professional responsibility for all evaluation methodology, analysis, findings, and recommendations.

• Assess risks, including potential bias, data privacy, and impact on any vulnerable groups, before and during AI use in evaluation work.

• Apply verification processes proportionate to the role AI plays in shaping findings.

Quality, accuracy and data

I carefully check all outputs from Claude and other AI tools for accuracy and quality. Information and data I share with Anthropic never includes personally identifiable or sensitive data. Additionally, data I share with Anthropic is never used to train or improve its AI models.

What I do not do

I seek to have boundaries and limits on my use of AI.

• I do not use AI to replace professional judgement.

• Present AI-generated content as original thinking without appropriate review and revision.

• Share personally identifying or sensitive data without fully informed consent and safeguards in place.

Questions or concerns

If you have questions or concerns about my use of AI, I am always happy to discuss them. Your trust is essential to our work together.

Last updated: June 2026