Study smarter with practical nursing learning resources.

Explore student support
OnlineNursingOwl

Structured prompt

Structured prompt

MY GRADUATE PROGRAM.....PSYCHOLOGY:

This assignment is open to either concentration. Each list below has six tasks. Choose three. For each chosen task, set evaluation criteria before generating, write a one-line baseline prompt and a structured prompt, run both, and iterate the structured prompt to at least version three. Counseling tasks • Draft a client psychoeducation handout that explains the cognitive model of anxiety in plain language, then ask the AI to flag clinical jargon and check the reading level. • Draft a plain-language explanation of informed consent for AI-assisted note-taking, for a new client. • From a short intake note, draft a referral letter to a psychiatrist, after defining what a good referral must include. • Draft three audience-tuned versions of a session summary: one for the clinician, one for the client, one for a referring provider. • Draft grounding-exercise instructions for a client, then ask the AI to flag any wording that overstates clinical benefit. • Draft a short front-desk script explaining the practice's AI-use policy to clients. I-O tasks • Draft a structured-interview question set for a named role, then ask the AI to flag any question that risks disparate impact or touches a protected characteristic. • Draft a candidate-facing FAQ explaining how AI is used in hiring and what human review applies. • From a fictional job description, ask the AI to audit for gendered language and inflated requirements. • Draft three audience-tuned versions of an interview-debrief summary: hiring manager, HR business partner, and a candidate-experience recap. • Draft a plain-language AEDT notice for candidates under a named AEDT law, then check it against the law's notice requirement. • Draft talking points for an HR leader explaining an adverse-impact finding to a hiring team.The same AI tool can give you a useless answer or an excellent one, and the difference is almost always the prompt. This assignment teaches you to build prompts deliberately, to decide what a good answer looks like before you ask, and to judge the answers you get instead of accepting them.

Your Task

Find your graduate program in the attached Program Task Briefs PDF and choose three tasks from your program's task list. For each of the three tasks:

  1. Before you generate anything, write down three standards the output must meet: what a good answer looks like, and what failure looks like.
  2. Write two prompts: a one-line baseline prompt (the quick version most people would type), and a structured prompt that includes an instruction, context, the format you want, and an example.
  3. Run both prompts. Copy both answers word for word.
  4. Score both answers against your three standards.
  5. Improve the structured prompt at least twice more (versions 2 and 3), writing down what you changed each time and why.

What to Submit

Submit one .docx or .pdf file containing the following sections, in this order:

  • Task Briefs: the three tasks you chose, described in your own words, 80 words or fewer each.
  • Evaluation Criteria: for each task, the three standards you wrote before generating anything.
  • Prompt Versions: every prompt word for word: the baseline, then versions 1, 2, and 3 of the structured prompt, for each task.
  • Output Samples: the AI's answer, copied word for word, for the baseline and for the final version of each task.
  • Grading Tables: for each task, a table scoring the outputs against your standards, with brief comments.
  • Iteration Notes: what you changed from version 1 to 2 to 3 and why, 300 words or fewer in total.
  • Reflection on Prompt Design: 300 words on what you learned about writing prompts.
  • AI-Use Statement: 150 to 250 words answering the four questions below.

The Four AI-Use Questions

  1. Which AI tools did you use, and at which steps?
  2. What prompts did you type? Include them word for word, or say where they appear in your document.
  3. What did you change in the AI's output, and why?
  4. How did you check that the facts are right? Name at least one thing you checked yourself against a source