Part 1, Understanding AI agents
Part 3, Introducing an agent
Appendices
- AA completed design record
- BChecklist
- CGlossary· You are here
- DReferences
- ERelated frameworks
Appendix C
Glossary
Scope and limitations of this guide
This guide provides general information on organising work with AI agents, not advice tailored to a particular organisation, system or decision. It is not legal or regulatory advice.
Its examples, checklists and templates do not establish that a proposed use is safe, lawful or suitable. You need to assess and test the proposed use in your own circumstances. Following the guide does not by itself meet the requirements of any law or standard.
The guide is provided in good faith, using information available at the time of publication or revision. AI systems, provider terms and legal requirements can change. Before acting, check the current requirements that apply to your organisation and country, confirm what the system can do and test the proposed use. You are responsible for how you apply the guide and should obtain professional advice where appropriate.
Reading, downloading or using the guide does not by itself create an adviser–client relationship with Tom Clowry or Considered Ideas Ltd.
To the fullest extent permitted by law, Tom Clowry and Considered Ideas Ltd accept no liability for loss or damage arising from use of, or reliance on, this guide. Nothing in this notice excludes or limits liability for death or personal injury caused by negligence, fraud or fraudulent misrepresentation, or any other liability that cannot lawfully be excluded or limited.
Completing this record or checklist does not establish that a proposed use is safe, lawful or suitable. Assess and test the proposed use in your own circumstances.
These terms are used in the guide.
Software that uses an AI model and available tools to work towards a goal over several steps.
The part of an agent that reads, reasons and writes, trained on large amounts of text and other material.
A defined way for one piece of software to connect to another.
Permission to make a decision or take an action.
The instructions given to the agent, including the purpose, permitted information and actions, checks and arrangements for seeking advice.
Examining whether the assumptions, reasoning or proposed approach should change.
The information and guidance an agent uses, including explanations of what the information means.
The amount of information a model can work with at one time.
An assessment of the risks to people of a use of their personal information, required where that use is likely to be high risk.
Giving an agent part of the work to carry out, with someone responsible for deciding the terms and overseeing the result.
A written account of how people and agents will share the work, how it will be checked and how the arrangements will be introduced and reviewed.
A systematic assessment of an AI system’s performance on representative tasks, using expected results or agreed criteria.
Controls that limit what an agent can do, such as access restrictions, approval steps and limits enforced by other systems.
A false statement, source or figure produced by a model as though it were true.
The working structure around AI models, which sets the stages of work, the tools, what is saved and checked, and when the task is complete.
The practice, feedback and experience people need to develop and maintain their expertise.
Information that an agent keeps and uses beyond the current request, such as saved notes.
An open standard that defines how an agent can use external tools and data.
A model that can work with images, sound or documents as well as text.
A model whose learned parameters, or weights, are available for others to download and run under its licence.
An attempt to redirect an AI system through instructions that override or conflict with those it should follow, including instructions embedded in material it is meant to read or process.
The improvement the work is intended to achieve and the people it should benefit.
A model trained to use additional computation to work through a task before answering.
Finding information in a source, such as documents or a database, and making it available to the model.
Examining the work and its evidence to establish whether it is reliable enough to use.
Permission given in advance for the agent to repeat a tested action under stated conditions.
An agent to which another agent passes part of a task.
A defined action that an agent can choose to take, such as searching, calculating or sending a message.
© 2026 Considered Ideas Ltd. Licensed under the Creative Commons Attribution 4.0 International licence (CC BY 4.0). This guide provides general information and is not legal or regulatory advice. The guide’s home page gives the suggested credit.
Scope and limitations of this guide
This guide provides general information on organising work with AI agents, not advice tailored to a particular organisation, system or decision. It is not legal or regulatory advice.
Its examples, checklists and templates do not establish that a proposed use is safe, lawful or suitable. You need to assess and test the proposed use in your own circumstances. Following the guide does not by itself meet the requirements of any law or standard.
The guide is provided in good faith, using information available at the time of publication or revision. AI systems, provider terms and legal requirements can change. Before acting, check the current requirements that apply to your organisation and country, confirm what the system can do and test the proposed use. You are responsible for how you apply the guide and should obtain professional advice where appropriate.
Reading, downloading or using the guide does not by itself create an adviser–client relationship with Tom Clowry or Considered Ideas Ltd.
To the fullest extent permitted by law, Tom Clowry and Considered Ideas Ltd accept no liability for loss or damage arising from use of, or reliance on, this guide. Nothing in this notice excludes or limits liability for death or personal injury caused by negligence, fraud or fraudulent misrepresentation, or any other liability that cannot lawfully be excluded or limited.

