Guide for tutors and teachers

What the toolkit can help with

The toolkit can help students use AI for focused writing support: feedback, explanation, questioning, revision planning and practice. It is most useful for low-stakes or formative work, where the aim is to help students understand how to improve.

For example, students can use it to:

• find repeated writing mistakes;
• understand why a paragraph is unclear;
• check whether an argument is too descriptive;
• turn tutor feedback into a revision plan;
• practise explaining a research proposal;
• record how they used AI during their work.

Set clear rules before students use it

Students need to know what is allowed in your course or setting. Before recommending the toolkit, explain:

• whether AI use is allowed for the task;
• what students may paste or upload;
• whether they need to declare AI use;
• which parts of the work must be done without AI help;
• who they should ask if they are unsure.

If your institution or course has its own AI policy, that policy comes first.

Be careful with private or sensitive material

Students and staff should avoid pasting private, sensitive or confidential information into public AI tools unless they are sure it is allowed.

This includes names, student numbers, email addresses, interview transcripts, placement notes, client details, unpublished research, marked work, feedback records or anything involving other people.

For classroom demonstrations, use invented examples or anonymised extracts.

Recommending it or using it in class or workshops

This is useful to teach directly. When students expect imperfection, they are more likely to question feedback, keep control of their own work, and check anything factual — especially references and sources — against course guidance. A student who treats AI output as provisional is already using it more maturely than one who treats it as authoritative.

Help students understand that AI tools are not predictable software. The same prompt can produce different answers, and the AI can be confidently wrong, drift off the chosen tool, or start rewriting work it was only asked to diagnose. This is a property of current AI tools, not a fault in the toolkit, and the toolkit reduces but cannot remove it.

Set honest expectations about the AI itself

How to recommend it to students

If you recommend the toolkit to students, make clear that it is a structured way to use AI for feedback, revision and learning support. Do not present it as a replacement for your course rules, university policy or tutor guidance.

You might say:

If you are using, or considering using, AI as part of your study or writing process — and I know many students are — please consider using the AI Personal Tutor Toolkit. It gives you a more structured way to use AI for feedback, revision and learning support, rather than asking AI to produce work for you.

You still need to follow the university’s AI-use guidance for your course, which is available here: [link]. If you are unsure whether a particular use is allowed, check with your module tutor before using AI on assessed work.

This wording keeps the emphasis on learning support. It also avoids implying that the toolkit overrides local assessment rules or has formal institutional approval unless that has actually been given.

A shorter version for class:

If you are going to use AI, use it in a structured way. This toolkit is designed to help you get feedback, understand problems and revise your own work. It is not for getting AI to produce work for you. You must still follow the AI rules for this module.

Using it in class or workshops

The toolkit can also be used in a live teaching session. For example, a tutor might project an invented paragraph, run one tool with the class, and ask students to judge whether the feedback is useful.

This works best with invented examples, anonymised extracts or short practice texts. Avoid putting identifiable student work, marks, feedback records or sensitive material into a public AI tool during a class demonstration.

A useful workshop pattern is: run one tool, discuss the feedback, ask students what they would accept or question, then ask them to revise their own work privately.

Use the toolkit as a conversation starter

AI feedback should not be the final word. A good use of the toolkit is to help students arrive at a better conversation with a tutor, supervisor or peer.

For example, a student might bring:

• the feedback they received from the AI;
• the revision plan they made from it;
• one point they agree with;
• one point they are unsure about;
• one change they made as a result.

The small, focused chunks the toolkit asks for are the working method, not a limit on ambition. Encouraging this repeated, developmental use also gives students a richer AI-use record to reflect on and, where required, to declare.

• at the planning stage, to test whether an idea, question or brief is workable;
• during drafting, to check one section at a time;
• after receiving feedback, to turn comments into a staged revision plan;
• near submission, to tighten writing they have already produced themselves.

The toolkit is most powerful when students return to it across a project rather than using it once. You can frame it to students as something to use at every stage:

Recommend it for the whole life of a piece of work

Start with low-stakes use

If you are introducing the toolkit to a class, start with a low-stakes activity rather than a live assessed task.

A safe first activity is to give students a short practice paragraph, ask them to use one tool, and then discuss whether the feedback was helpful, too vague, too strong, or too close to rewriting.

This helps students learn how to question AI feedback rather than simply accept it.

Tell students they can say “I’m stuck”

Students do not always know which tool they need or what kind of help to ask for. It is useful to tell them that they can say “I’m stuck” at any point.

The tutor should then step back, offer a small number of possible ways forward, and ask whether it has understood the problem. This is often more useful than giving students a long menu of every possible issue.

Tell students they can disagree

AI feedback is not authoritative. Students should know they can correct the AI, explain what they meant, ask it to reread their work, or say that a suggestion does not fit their meaning.

If they are unsure whether the feedback is right, they can use it as a question to take to a human tutor, supervisor or peer. The aim is not to obey the AI, but to use it to support better judgement and better revision.

What library to use

Most students do not need the full master library. It is usually better to use the smallest library that fits the task.

• For sentence-level writing help, use the Writing Tutor Library.
• For paragraph and whole-draft structure, use the Structure Tutor Library.
• For argument, evidence and concepts, use the Academic Thinking Tutor Library.
• For dissertation or project planning, use the Research Proposal Tutor Library.
• For revision planning and AI-use records, use the Study Workflow Tutor Library.

The master library is useful when students need access to everything, but smaller libraries can make the AI easier to steer. If students are using free AI accounts or a mini library is still too large, use the single-tool downloads as a smaller backup route.

When to use the testing pack

If you are recommending the toolkit to a group, it is worth checking how it behaves in the AI tool your students will use.

The testing pack gives sample checks for whether the toolkit stays within its teaching role. It can help you spot problems such as over-writing, unclear explanations, false corrections or weak handling of student pushback.

You do not need to test everything before a small informal use, but testing is useful before a wider rollout.