Module 01 · Your AI Job-Search Toolkit
AI-Assisted Job Search: Where It Helps and Where It Does Not
Open lesson + course map
On this lesson
Course outline
Module 1 · Your AI Job-Search Toolkit
Module 2 · Resume Engineering
Module 3 · LinkedIn as a Lead Engine
Module 4 · Interview Preparation With AI
AI is useful when a job-search task involves sorting information, drafting, or practising a conversation. It becomes risky when you ask it to judge your chances, speak for an employer, or fill gaps in your experience.
By the end, you will have a one-page rule sheet showing where you will use AI, where you will not, and what evidence you must check before sending an application.
// concept
Map the Task Before Using AI
Score the task, not the tool. A chatbot can organise facts you provide; it cannot verify your memory or reveal an employer’s private decisions.
| Job-search task | Score | Reason |
|---|---|---|
| Extract requirements from a job post | Helps | It can organise explicit requirements; check against the original. |
| Compare a redacted resume with a job post | Helps | It can surface matches and gaps without deciding your eligibility. |
| Rewrite a truthful resume bullet | Helps | It can improve clarity if every detail remains verifiable. |
| Draft a role-specific cover letter | Helps | It can provide structure; you must verify company facts. |
| Generate mock-interview questions | Helps | It can turn the public description into practice. |
| Research an employer | Neutral | It suggests searches; current facts need official, dated sources. |
| Estimate salary | Neutral | It can organise sourced ranges, not invent a “market rate.” |
| Submit many applications automatically | Hurts | Automation can miss instructions or send the wrong resume. |
| Predict an ATS score or interview chance | Hurts | The model cannot see the employer’s configuration, applicant pool, or review process. |
| Add missing experience, metrics, or credentials | Hurts | A polished fabrication is still a false claim and may fail verification or interview questions. |
Ethics is simple: never use AI to fabricate experience, qualifications, titles, dates, references, or results. Mark a missing requirement as a gap; learn it, demonstrate it honestly, or skip the vacancy.
// concept
Give AI Evidence, Boundaries, and an Output Format
Use a three-part input: the job post, a redacted resume, and explicit rules. Remove your CNIC, full address, private phone number, references’ details, and confidential employer or client information first.
You are helping me review one job application.
JOB POST:
[paste the dated job description]
MY VERIFIED EVIDENCE:
[paste a redacted resume or selected project notes]
TASK:
Create a table with: Requirement | Evidence I supplied | Status | Next action.
Use only these statuses: Supported, Related, Gap, Unclear.
RULES:
- Quote the relevant job-post phrase in the Requirement column.
- Do not invent or infer experience, tools, dates, metrics, qualifications, or employer preferences.
- If my evidence does not prove a requirement, mark Gap or Unclear.
- Do not calculate an ATS score or predict an interview.Check every “Supported” row. Turning “helped with weekly reports” into “owned reporting,” or a class exercise into work experience, changes the facts.
// concept
Recognise AI-Shaped Applications Without Chasing Detectors
A recruiter may notice generic writing that repeats the advert or collapses under interview questions. That does not prove AI authorship. Claims that recruiters “always detect AI,” and detector scores presented as proof, are not sound guidance.
Use the say-it-aloud test: “Can I explain this naturally, name its evidence, and answer a follow-up?” Replace “results-driven professional leveraging synergies” with the work, tool, audience, and constraint you can defend.
// worked_example
Worked Example
This is a hypothetical sample. A Lahore fresh graduate reviews a sample Junior Data Analyst post requiring Excel pivot tables, basic SQL, dashboards, and stakeholder communication. Their evidence: a university survey cleaned in Excel, two pivot summaries, an introductory SQL course, and a team presentation—but no dashboard.
They use the prompt above. A plausible first-draft excerpt is:
| Requirement | Evidence supplied | Status | Next action |
|---|---|---|---|
| Excel pivot tables | Two pivot-table summaries for a university survey | Supported | Add the project and explain the analysis question |
| Basic SQL | Introductory SQL course | Related | Link a small query project if available |
| Dashboard experience | University survey analysis | Supported | Describe the dashboard created |
The third row is wrong: no dashboard was supplied. The candidate fixes the input with this instruction:
Correction: I did not build a dashboard. Re-evaluate every row. “Related” means adjacent
experience; it does not permit you to rename an activity. Quote my exact evidence before
assigning a status, and mark absent evidence as Gap.The corrected row is “Dashboard experience | No evidence supplied | Gap | Build a portfolio dashboard or do not claim this skill.” The value is a truthful next action, not a higher score.
// failure_cases
Failure Cases to Diagnose
6 cases to diagnose
The table marks silence as evidence
Require a quotation from your resume for every supported match.
A course becomes work experience
Keep training under education or projects unless paid or employed work actually occurred.
The draft copies the job post
Add one concrete example from your evidence instead of repeating the employer’s wording.
The model assigns an ATS percentage
Delete it; the model does not know the employer’s filters or recruiter workflow.
One cover letter fits every company
Recheck the role, employer facts, motivation, and evidence.
A confident company claim has no dated source
Verify it on the employer’s official website or remove it.
// pakistan_angle
Pakistan Angle
For a Pakistani search, note whether the source is an employer careers page, LinkedIn, Rozee.pk, Indeed, or a forwarded WhatsApp post. Follow the listed channel. LinkedIn distinguishes Easy Apply from Apply, which may redirect externally; check where your data goes. Stop and independently verify early requests for a CNIC scan, bank information, or payment.
On mobile data or during load-shedding, save the post and redacted text resume locally. Draft offline, then send one compact prompt. Keep city, onsite availability, notice period, and English/Urdu requirements accurate. For USD remote roles, verify contractor status, taxes, hours, and payment rails; do not let AI present them as settled.
// hands_on
Hands-On Exercise
6 steps
Build a “Use AI for / Never use AI for” one-pager:
Copy the ten-task table and change each score to match your workflow.
Under “Use AI for,” select three tasks and write the evidence you will provide for each.
Under “Human review required,” list facts you will verify: employer, vacancy date, claims, attachment, and submission channel.
Under “Never use AI for,” include fabrication, autonomous submission, and unsupported predictions.
Add your redaction list: CNIC, address, phone, reference details, and confidential work data.
Test the sheet on one current job post; correct any AI output that exceeds your evidence. “Done” means the one-pager can guide a real application without relying on memory or a detector score.
// completion_rubric
Completion Rubric
5 checks — tick as you verify
// sources
Sources
3 official sources — check every claim yourself
// check_yourself
Check yourself
4 questions · answers and options are taken word-for-word from this course
1 / 4 · diagnose
Your work shows this failure mode: “The table marks silence as evidence.” What does the lesson tell you to do about it?