Gains:
- Ability to connect a weekly PDR workflow to an end-to-end process supported by artificial intelligence and measure time savings
- Ability to create a reusable prompt library and personal PDR assistant instructions appropriate to one's own professional context
- Ability to link verification, confidentiality and crisis boundaries to a checklist throughout the entire process and turn it into a permanent habit
In the final unit, we combine everything we have learned into a single whole. So far, we have used AI separately in every professional field, from interview notes to psychoeducational materials, from career guidance to scale interpretation support, from follow-up to group activity; We covered limits, authentication, privacy and crisis exception in each area. Now we will tie these into a weekly workflow, create a context-specific prompt library and personal PDR assistant instructions, and secure the entire process with a permanent checklist. The goal is to make AI a safe and permanent part of your professional life, not a one-time gimmick.
Designing a week from end to end
A typical week for a PDR professional includes a lot of repetitive writing work. Putting them into an AI-supported flow saves time and ensures consistency. An example week:
- Monday — Planning. The draft structure of the week's group guidance sessions (acquisition–warm-up–main–closure–evaluation) is generated by AI and subjected to expert security and comprehensiveness review.
- Tuesday — Material. A draft psychoeducational brochure is produced; statistics are marked with [VERIFY] and confirmed.
- Wednesday — Interview and registration. Individual interviews are held; Afterwards, anonymous raw notes are structured with AI, and comments and decisions come from the expert.
- Thursday — Collaboration. Drafts of parent information letters and teacher collaboration notes are produced and subjected to tone and confidentiality checks.
- Friday — Follow-up and recap. Anonymous follow-up data is summarized, patterns are evaluated as signs, and the reasons for next week's meeting are determined.
In this flow, AI always plays the same role: drafting and editing. Decision, interpretation, security and final approval are all in the hands of the expert. Crisis is always beyond the threshold; When a sign of risk occurs, the flow stops and the protocol in unit 9 is activated.
Measuring time savings
To see the value of a habit, it is necessary to measure the gain. For a week, keep a note of "how long it took before / how long it takes now" for each task you speed up with AI. Most experts see a 40-60% time savings on repetitive typing. But the real question is: Where did you devote the saved time? The correct answer is not “more paperwork” but “more face-to-face time with the student.” The goal of AI is to get you off the table and back to the human essence of the profession: the relationship with the student.
Tip: Record your time savings in a table: task, old time, new time, gain, where you allocate the gain. This table both motivates and shows where AI can be used more.
Personal prompt library and assistant instruction
Instead of writing prompts from scratch every time, save the prompts that work in a library. A good prompt library; It consists of prompts arranged according to task type, tested, and with confidentiality and verification limits coded into them. In addition, you can define a personal assistant directive (system instruction) in your favorite tool: permanent rules that the tool will follow every time it talks to you.
Good PDR assistant guidelines include: role (writing assistant assisting the PDR specialist), what they will not do (no diagnosis, no referral, no role in crisis), what they will always do (separating interpretation from fact, assumption of anonymity, verification warning, crisis reflex reminder), tone (empathetic, non-stigmatizing, solution focused). This directive hardcodes security into every interaction.
Copiable templates
1) Personal PDR assistant instruction (system instruction):
You are a WRITING assistant to a school counselor. What you will NEVER do: make a diagnosis, make decisions/directions on behalf of the student, interpret a scale, suggest intervention in a crisis situation, reduce the student to a label. What you will ALWAYS do: separate fact from interpretation; assume the input is anonymous and warn you if you see personal data; putting a "VERIFY" note next to printouts containing facts/statistics/sources; If you sense a sign of crisis/risk, write the warning at the top: "This may be a crisis; not AI, but human expert and emergency protocol should be activated." Tone: empathetic, non-stigmatizing, solution-oriented.
2) Weekly time savings table:
Make a table for the following tasks: task, time before AI, time with AI, time saved, where I separate the gain. Also calculate total earnings.Tasks and durations: [enter your own data]
3) Prompt library editing:
Categorize my scattered prompts below by task type (recording, material, career, follow-up, collaboration, group) and add a brief usage note for each category. Changing their content.Prompts: [paste prompts]
4) End-to-end verification control:
Before publishing/using the AI output below, apply these 5 checks and write “ok/problem” for each: (1) personal data leak, (2) unverified fact/statistic/source, (3) diagnosis/label/referral, (4) bias/cultural insensitivity, (5) crisis boundary violation. Flag issues.Output: [paste output]
Weak prompt / Strong prompt
Weak:
Help me with my work this week.
Contextless, unlimited; Which task, which role, which boundary is unclear — it turned out to be general and risky.
Strong:
[Personal PDR assistant instructions are active.]Task: This week I need to plan a "friendship skills" group session for 6th graders. A draft is created in the structure of gain-warm-up-main event (fictional scenario)-closing-exit card. Also mark the points you deem risky in terms of security and comprehensiveness. If there is a claim that needs to be verified, put VERIFY.
three mini cases
Case 1 — Measured gain. An expert recorded the tasks he accelerated with AI in a table for a month. He found himself saving ~6 hours a week total in record writing, material preparation, and collaboration letters. He devoted this time to the individual meetings he had previously lined up; The number of meetings increased and student satisfaction increased. The profit returned to the student, not to the table.
Case 2 — Safety by directive. A guidance counselor absent-mindedly left a student's name in a text one day after describing personal assistant instruction. Thanks to the "warn if you see personal data" rule in the directive, the tool reminded this before starting the process. The expert pulled out the name. The guideline served as a safety net against human error.
Case 3 — Consistency with the library. One service collected its scattered requests into a common request library; everyone started using the same, security-limited coded prompts. Instead of a weeks-long learning curve, a new colleague went into production safely on day one using the library. Information moved from person to institution.
Common mistakes
- Unlimited requests. Saying "help" without context. Solution: clear role, task, format, boundary.
- Not defining a directive. Manually setting security every time. Solution: permanent assistant directive.
- Not measuring the gain. Not seeing improvement. Solution: time savings table.
- Misappropriating profits. Spending the time saved on the table again. Solution: return time to the student.
- Forgetting to verify at the end of the flow. Relaxing control as speed increases. Solution: end-to-end authentication control.
Table: Weekly flow and assurance
day
business
AI role
Expert assurance
monday
Session plan
draft
Security/coverage audit
tuesday
material
draft
Statistics verification
Wednesday
Registration
Editing
Fact/interpretation separation, anonymity
Thursday
collaboration
draft
Tone and privacy control
friday
tracking
Summary
Cue→interview, bias control
every moment
crisis
None
Human protocol (unit 9)
In summary
Throughout this module, we learned to use AI as a writing and drafting assistant in every aspect of PDR, maintaining verification, confidentiality and crisis boundary at every step. The final step is to plug them into a weekly flow, set up a personal library of prompts and assistant instruction, measure the time savings, and make an end-to-end verification check a habit. The goal of AI is to get you off the table and back into your human relationship with the student; The decision, interpretation, security and responsibility always remain with you.
Application task
Write your own weekly PDR workflow. Put each repetitive writing task on one line and specify the AI's role (draft/edit/summary) and your assurance (authentication/privacy/security). Then adapt the “personal PDR assistant instruction” template to your context and create a starter library of at least 5 prompts. Finally, measure your time gain for a week and write down where you allocated the gain.
checklist
- [ ] I paired my weekly workflow with the AI role and expert assurance.
- [ ] I created a personal assistant instruction and prompt library.
- [ ] I measured the time savings and directed it towards time spent with the student.
- [ ] I made end-to-end verification control (5 items) a habit.
- [ ] I have ensured that the crisis limit is valid at every moment of the flow.
Module Exam
1. A school counselor is considering asking an artificial intelligence tool how to handle suicidal thoughts expressed by a student in an interview and implementing the response. What is the right approach?
- A) In case of a crisis, immediately operate the emergency protocol instead of asking artificial intelligence to make a decision; ✔ Applying to competent human specialist, family and necessary authorities
- B) Implementing the intervention plan given by artificial intelligence as it is
- C) Give the artificial intelligence the full name of the student and request a personalized plan
- D) Record the conversation and have the artificial intelligence diagnose
Explanation: Crisis and risk situations (suicidal thoughts, harm to self/others, abuse, neglect) cannot be transferred to artificial intelligence. In these cases, it is mandatory to immediately contact the competent human expert, family and necessary authorities and operate the emergency protocol. Artificial intelligence can be used mostly in post-crisis record editing, never in decision making and intervention.
2. What is it called when artificial intelligence produces 'fluent but actually false' information (a made-up statistic, a non-existent source, a wrong scale norm) and what should a PDR expert do?
- A) Hallucination; verifying every fact, statistic, and source from a reliable source
- B) Hallucination; relying on direct because the output is fluent
- C) Hallucination; ✔ Verify every output containing facts, statistics and sources from a reliable source and use it accordingly
- D) Bias; not using the output at all
Description: Artificial intelligence predicts and produces the possible word; Therefore, it can give fluent but incorrect information (hallucinations). Any output that reaches the student, such as psychoeducational material, statistics and resources, should not be used without being verified and linked to a reliable source by an expert.
3. A guidance counselor wants to consult the artificial intelligence about a difficult student situation. What is the first thing he should do in terms of privacy?
- A) Write the student's name and family and get a personalized response
- B) Add the parent's phone number and ask for contact suggestions.
- C) Paste the conversation record as it is
- D) Anonymizing text: consulting after replacing name, number, class and identification information with general expressions ✔
Explanation: Student data is personal data and psychological/family status is special personal data and is protected within the scope of KVKK and professional confidentiality. Writing credentials into a public tool puts data out of control. Rule: anonymize first; Replace information such as name, number, class, etc. with general expressions such as 'Student A', 'A 13-year-old student'.
4. Which of the following is the most accurate positioning of the use of artificial intelligence in career guidance?
- A) Making final career decisions on behalf of the student
- B) Producing questions and information drafts that facilitate student discovery; Leave the decision to the student and verify the data from the official source ✔
- C) Getting the current score and quota from artificial intelligence and telling it without verifying it
- D) Assigning a student a profession without making an interest inventory
Description: Artificial intelligence is used as a knowledge outline and discovery question generator in the career discovery and selection process; A discovery-oriented approach that centers on the student's interests, values and competencies is supported. There is no clear 'choose this profession' guidance; Information such as scores and quotas are verified from official sources.
5. What is one of the most important principles when using AI when writing an interview transcript?
- A) Allowing artificial intelligence to put diagnosis-like labels on the student
- B) Basing the record solely on interpretation
- C) Keeping an objective, observational, anonymized and expertly verified record that separates fact from interpretation ✔
- D) Add identification information to the record and use the tool as an archive
Explanation: A good recording separates fact (observed, said) from interpretation (expert's inference). AI can help with note organization but carries the risk of tagging and commenting; The record must be unbiased, observational, anonymized and expertly verified.
6. What should be the role of artificial intelligence in scale and inventory results?
- A) Give raw scores to artificial intelligence and request clinical interpretation and diagnosis
- B) Use as an editing assistant to improve the presentation and language of expert commentary ✔
- C) Adapting norm tables to artificial intelligence
- D) Having the student explain the result to artificial intelligence
Description: Scale interpretation, diagnosis and clinical meaning; It requires validity-reliability and norm knowledge and belongs to the expert. At best, AI can be an editing/language assistant that translates anonymized, expertly interpreted results into understandable language; cannot interpret the result itself.
7. What is the most accurate limit for an expert who summarizes absence and observation data with artificial intelligence for early warning purposes?
- A) Turning the 'risky' label of artificial intelligence into a direct decision
- B) Opening the list to share with student names
- C) Just looking at the data and not meeting with the student at all
- D) Taking the summary as a signal and combining it with expert observation and interview; ✔ Leave the decision to the expert
Explanation: AI can produce summaries that make it easier to spot patterns, but it can reduce the learner to a 'risk' and introduce bias. Early warning should always be combined with the specialist's observation and discussion with the student; The final evaluation and decision should belong to the expert.
8. What is the most beneficial use of artificial intelligence in drafting a challenging parent meeting letter?
- A) Producing an empathetic, non-blaming and solution-oriented communication outline; Arranged by the expert according to confidentiality and student interest ✔
- B) Adding all sensitive information of the student to the text to be sent to the parent
- C) Using harsh language that blames the parent
- D) Send the draft as is without reading it at all
Description: AI is good at creating and drafting language that is empathetic, non-blaming, and solution-focused. The expert prepares this draft, taking into account confidentiality and the student's benefit; It anonymizes personal data and gives final approval.
9. What is the most critical check that the expert should make before applying the AI output in the group guidance session plan?
- A) Whether the activity is given with a colorful presentation or not
- B) Writing student names on the activity plan
- C) Making the activity duration as long as possible
- D) Expert evaluation of the activity in terms of emotional safety, inclusiveness and compliance with classroom reality ✔
Explanation: Activities suggested by artificial intelligence must pass expert scrutiny in terms of emotional safety, inclusiveness and feasibility. Even though an activity may seem good in theory, it may be invasive or exclusionary in a certain class; This judgment belongs to the expert.
10. Which of the following falls within the scope of 'special categories of personal data' and requires higher protection?
- A) Information about the student's mental state, diagnosis and family situation ✔
- B) General curriculum of the school
- C) General text of a psychoeducational brochure
- D) A public profession introduction letter
Explanation: According to KVKK, information such as health, mental status, family situation, diagnosis, sexual life are special personal data and require higher protection. Most PDR records fall within this scope; That's why public vehicles should never be accessed in their raw form.
11. Why is the statement 'AI said' not a valid justification for a PDR expert?
- A) Because artificial intelligence is always correct
- B) Responsibility and final decision belong to the expert; Unverified output is not a substitute for professional judgment ✔
- C) Because artificial intelligence should never be used
- D) Because artificial intelligence can only be used by managers
Disclosure: AI output is not a substitute for the professional judgment and responsibility of a competent expert. The expert has the responsibility and the final say in every decision that affects the student's future, psychological well-being and rights; An unconfirmed report is like a rumor whose source has not been verified.
12. What should be done before using an AI statistic ('X% of young people experience test anxiety') when producing a psychoeducational brochure?
- A) Match the source to artificial intelligence and add it
- B) Putting it directly on the brochure because it looks fluid
- C) Enlarge the number to make it more impressive
- D) Verifying statistics from a reliable scientific/official source, not using them if they cannot be verified ✔
Description: Artificial intelligence can fabricate statistics and sources (hallucination). A scientific claim that will reach the student and the parent should not be used without being verified and linked to a reliable source (scientific study, official institution data); Otherwise, false information will spread.
13. Which of the following is true when PDR professional ethics principles (confidentiality, beneficence, non-maleficence, autonomy) are adapted to the use of artificial intelligence?
- A) Ethical principles are valid only in face-to-face meetings
- B) The use of artificial intelligence automatically fulfills ethical principles
- C) Ethical principles are also binding in the use of artificial intelligence; Confidentiality, non-maleficence and autonomy must be protected under expert responsibility ✔
- D) Ethical principles have lost their validity in the age of artificial intelligence
Explanation: Ethical principles also apply in the use of artificial intelligence: confidentiality requires anonymization, non-maleficence requires verification and crisis limit, autonomy requires not reducing the student to a label. Artificial intelligence is a tool that can violate these principles, not facilitate them; The responsibility lies with the expert.
14. What statement best summarizes the fundamental difference between a weak will and a strong will in the context of PDR?
- A) Strong will is always better because it is longer
- B) The strong request makes the role, context, format and boundaries (anonymity, crisis, verification) clear; weak claim is vague and lacks context ✔
- C) Weak prompt is safer because it gives little information
- D) There is no practical difference between the two claims.
Description: Strong will; It clearly states the role, context (age, purpose), format, boundaries (anonymity, crisis boundary) and expectation of verification. The weak prompt is vague and contextless, producing risky and generic output. Good prompting encodes expert control and confidentiality from the start.