A sequence of prompts needs a shared purpose, visible assumptions and a clear place for the reader to review the result.
An original educational exercise examines how instructions and evidence move through a sequence of drafting steps.
Keep the purpose visible across each instruction
A longer task often contains several kinds of work. A reader might need to find a relevant passage, explain what it means and prepare a draft for someone else. In this fictional exercise, those activities are written as separate prompts. The point is not to make the sequence longer. It is to examine what each step receives, what it produces and how the result remains connected to the original request.
Begin with a synthetic note that describes a process and one unresolved condition. The first instruction asks for the passage relevant to a particular question. The second asks for an explanation limited to that passage. The third asks for a draft that preserves the unresolved condition. Each instruction has a different output, but all three serve the same defined purpose.
Describe that purpose in language another reader can understand. A request to make a useful summary leaves many choices unstated. A request to explain a named condition and prepare a question for review gives the sequence a clearer boundary. The expected result should identify what the reader needs next without pretending that the source has resolved something it leaves open.
Inspect what passes from the first step to the second. If only a shortened quotation moves forward, the explanation may lose the condition that gives it meaning. Carry enough context to distinguish the relevant passage from an example or an older instruction. The next step should not need to guess why the passage was selected or what question it was meant to answer.
Then inspect the transition from explanation to draft. An interpretation should remain identifiable as an interpretation. The final wording should not quietly turn a possibility into an instruction or omit the question that still requires review. A reader comparing the draft with the source should be able to follow the connection and recognize where judgment was introduced.
Try a variation in which the source does not contain the requested detail. The sequence should make that absence visible at the point where it matters. Carrying an unresolved question forward can be more useful than filling the gap with plausible language. The exercise can ask whether each step preserves the uncertainty or whether a later instruction accidentally encourages it to disappear.
Another variation can introduce two conflicting synthetic notes. Ask the first step to retain both relevant passages and explain their relationship. The later draft should not resolve the conflict merely to produce a cleaner answer. Instead, it can name the condition that needs clarification. This gives the reviewer an explicit question rather than an invisible assumption embedded in finished-looking prose.
Save the instructions alongside the outputs when reviewing the exercise. If the wording changes between attempts, record that change. Otherwise, a reader may attribute a difference in the draft to the wrong cause. The purpose of the record is to make the discussion inspectable, not to report a general performance result from a small set of fictional examples.
A sequence is ready for discussion when another reader can explain the role of each step and inspect the final draft against the source. That does not establish that a real workflow is ready to deploy. It means the exercise has made purpose, evidence and assumptions visible enough to review. The next useful step may be to simplify an instruction, clarify the expected output or leave a decision unresolved.
About the prompting exercise
Original illustrative material. No funding, product release, customer adoption or productivity finding is claimed.
The notes, instructions and drafting sequence described here are fictional. The article provides an educational way to examine the handoffs within a multi-step task. It does not announce a prompting feature, investment, customer relationship or observed improvement.
The examples can be practised with synthetic material. Keep any conclusions limited to the particular instructions and outputs examined, and preserve the questions that the exercise does not answer.
Continue exploring
The original evaluation and prompting guides extend the exercise with questions about source scope, expected outputs and a readable review handoff.