A short introduction for those who want to use AI to support thinking, reflection, and preparation in their work. This post offers a simple way to shape prompts that keep human judgement, relationships, and context at the centre.
In an earlier post, I reflected on using AI as a quiet thought partner in my own work. This piece builds on that by offering a light structure for prompting that supports thinking, writing, and preparation while keeping human judgement and relationships at the centre. It is written for people who want to use AI tools in reflective and collaborative ways, without getting pulled into technical advice about prompt engineering.
Recent management writing points to a shift that matters for those of us working in complex, multi-actor settings. David Duncan and colleagues describe how AI now takes on much of the research, analysis, and early drafting that once sat with junior roles, speeding up the work yet still needing human judgement to shape outputs for real-world use. At the same time, commentary such as Kelli María Korducki’s suggests that the value of consultants, advisors, and facilitators is shifting, resting less on providing answers and more on helping people interpret insight, work through uncertainty, and apply ideas in practice.
This is where purposeful prompting can help. Rather than looking to AI as a tool to do the work for us, we can use it to support our own sense-making and preparation.
Why prompting matters for reflective practice
There is now plenty of introductory guidance on writing effective prompts for clearer or more polished outputs. The MIT Sloan short guide on effective prompts offers a useful introduction to prompt mechanics, and works well for task-based outputs such as summarising or drafting. For those working in reflective or collaborative settings, the value often lies less in polishing outputs and more in shaping the thinking that leads to them.
This is particularly relevant for those of us who use reflection, dialogue, and preparation to support collective learning and change.
A more helpful starting point is to treat prompting as a small act of intention. A clear prompt can shape the kind of thinking that follows. In complex or relational work, the value often lies not in speed, but in how well AI can help us explore, reflect, and prepare.
The light framework below has grown from practice rather than theory. It is not a set of rules, simply a way to invite AI into the thinking process without losing sight of the human work that matters. This hands-on approach aligns with other practice-informed frameworks in the field, including Mo Hoque’s Thinking with AI: A framework for expanding how you work, learn, and create.
Five everyday purposes for prompting
Most prompts fall into one of five everyday uses. Naming the purpose before you write a prompt helps you stay clear about what you are actually seeking from the tool. It means you are less likely to ask for a finished answer when what you really need is inspiration, sense-making, or help getting started. This small pause can shape the quality of the exchange and keep the AI in a supportive role rather than leading the thinking.
1. Explore
Useful when you want to widen perspectives, surface ideas, or scan what is out there. Explore prompts work well early on, when you are still shaping your understanding of an issue or deciding where to focus. They help open up a question, pull in fresh angles, or point to material you may not have considered. The aim is not to find a solution quickly, but to stretch your field of view before narrowing things down.
Example openings:
• Help me explore …
• What are some different ways people think about …
• Show me a few contrasting perspectives on …
2. Clarify and reflect
Helpful for making sense of notes, teasing out themes, or testing assumptions. Clarify prompts work well when ideas feel a bit tangled or unfinished. They can help you notice patterns, check what may be missing, or see how different people might interpret the same material. The aim is not to tidy everything up, but to slow the pace a little so you can reflect before moving forward.
Example openings:
• Here are my notes, help me surface the key themes …
• What assumptions sit behind this idea …
• Where might different groups see this differently …
3. Develop and create
Useful for early shaping. It can turn fragments into outlines, help reframe an idea, or offer options you can build on. Develop prompts are most helpful when you are ready to make progress but do not yet know the best form. They provide a starting structure that you can then adapt, rather than facing a blank page or relying on a single approach.
Example openings:
• Turn these bullet points into a rough outline …
• Offer three ways to frame this issue for discussion …
• Draft a short version I can refine …
4. Improve and adapt
Works well when you already have material and want to strengthen it. Improve prompts work best once you have your own thinking on the page. They help refine flow, tighten language, or check clarity, without handing over authorship. The aim is to enhance your work while keeping the voice, intent, and judgement firmly yours.
Example openings:
• Help me tighten this text without changing the meaning …
• Suggest two clearer ways to introduce this section …
• What might confuse readers here …
5. Prepare for engagement
Can be valuable when getting ready to work with others. Prepare prompts can help gauge how material may land with different audiences, anticipate questions, or shape inclusive language that supports participation. They are most valuable before a workshop, meeting, or publication, when a little rehearsal can help you show up with more care and clarity.
Example openings:
• How might this be read by …
• What questions might participants ask about …
• Suggest inclusive language for introducing this topic …
A light structure for a clear prompt
Alongside choosing a purpose, a little structure can make a prompt more effective. A helpful prompt often includes three simple elements. You do not need all of them every time but using one or two can make the exchange more purposeful. These elements give the model enough context to support your thinking, without over-steering the response or squeezing out your own judgement. Many people find that adding just a short line of context or a gentle boundary helps keep the tool in a supportive role rather than taking over the voice or direction.
1. Context – the situation you are working within
A short line about the situation, audience, or task. This helps the model understand the setting you are working in, so the response can be more relevant without needing lots of detail.
• I am preparing a 90 minute workshop with council staff on freshwater planning…
2. Request – the type of support you want
The specific help you want. This keeps the focus clear. Being direct about the type of support you are seeking often produces more useful ideas on the first attempt.
• Offer two possible ways to open the session that encourage reflection, not debate.
3. Preferences or boundaries – how you’d like the response shaped
Tone, format, or other constraints. These help guide the response so it aligns with your purpose, voice, or audience. They can be especially helpful when using AI to refine or draft written material.
• Keep the language plain. Avoid jargon. Use British spelling.
If you want the model to stay close to your own style, you can add a short sample of your writing. A paragraph is usually enough. Some people also find it helpful to name the perspective they are bringing, such as writing as a facilitator, evaluator, or team lead. This can help the model respond with a more relevant lens.
A short reflection from practice
In my own work, having a clear purpose and a light structure for prompting has helped me use AI in ways that support rather than shape my thinking. For example, when preparing a workshop, I often use the ‘explore’ and ‘clarify and reflect’ prompts, along with a short line of context and the perspective I am bringing as a facilitator, to check whether my session framing would resonate with different participant groups. The aim was not to produce answers, but to help me notice what might need more attention.
When drafting a resource page, I tend to combine ‘develop and create’ and ‘improve and adapt’ prompts with a brief boundary about tone. This helps me move from scattered notes into a more coherent flow, while still holding purpose and voice as I rewrite sections.
Ahead of sharing a piece more widely, I sometimes use ‘prepare for engagement’ prompts to test how it might be received by different audiences. Naming who I have in mind for the piece helps adjust language so it feels more inclusive and avoids unintended signals.
In each case, the tool supports thoughtful, constructive reflection, while the decisions remain mine.
Using AI in collaborative work
AI can help us organise thoughts, surface options, and prepare for conversations. It can speed up parts of the process, although faster is not always better. In collaborative work, the value lies less in efficiency and more in how the tool supports clarity, care, and shared understanding.
A few simple habits can help:
- Treat the model as a thinking companion, not an answer giver.
- Keep your own judgement at the centre.
- Use AI to broaden and structure ideas, then explore meaning with people.
- Be transparent about when and how AI was used.
- Avoid sharing sensitive or confidential material. Summarise if needed.
When not to use AI
There are moments when AI gets in the way. It cannot replace relationship building or real dialogue. It cannot sit with uncertainty or read a room. It cannot speak for lived experience or cultural ways of knowing. It struggles with context and the social side of change.
Recent research on collaborative sensemaking between humans and AI has explored these boundaries in more depth. For example, the recent paper by Xinyue Hao and colleagues, Beyond human-in-the-loop: Sensemaking between humans and AI, highlights where the tool can fall short. For decisions that need presence, listening, or shared meaning-making, it is often worth keeping the tool in the background.
Concluding comments
Prompts are not the main event. They are simply a way of bringing AI into the thinking process with intention and care. Used well, prompts can help us slow down, broaden inquiry, and strengthen how we show up in our work. The value lies not in producing perfect answers, but in supporting our own reflection and preparation.
For most of us working in complex, multi-actor settings, the central task is still to foster understanding and connection. AI can assist, yet the heart of this work remains human. It can help us arrive more prepared, more thoughtful, and more open to one another.
This post has focused on individual practice, and how you as a practitioner can bring AI into your reflective routines without losing judgement or voice. In a future piece, I will explore how teams and groups can use AI to deepen learning, surface diverse perspectives, and support shared ownership, so that collaboration stays at the centre rather than being displaced by automation.
More information on the use of AI can be found through the LfS generative AI landing page. This includes links to the related resource pages Using AI in research and practice and and AI in context: the wider picture. Also see the accompanying posts:
- AI as a thought partner: reflections on collaborative practice and systems work
- Working with AI in the room: authorship, responsibility, and collective judgement
- Seeing the wider ethical picture around AI development and use
- AI in place-based practice: what is shifting
- Working with AI: where to begin
Collectively these pages introduce practical tools, ethical guidance, and curated articles to support thoughtful use of AI in business, applied research, and collaborative practice.
[* Photo by:ThisIsEngineering | Pexcels]
