Artificial Idea | AI careers · practical prompts · no hype Thursday, October 23, 2025 · Issue #24 · Prompt Tutorial
The resume problem, solved
How to use AI to rewrite your resume for any job description in 10 minutes
Most resumes are written for the person who wrote them. The ones that get interviews are written for the person reading them. These prompts close that gap in ten minutes.
Issue #23 made the case that constraint-trained professionals have built cognitive capabilities that transfer directly to AI-augmented work, and that the professionals who recognise and build on those capabilities will outperform those who continue measuring themselves against standards designed for different contexts. This issue puts that argument into practice in the highest-stakes document most professionals produce: the resume.
The resume problem in 2025 has two distinct layers that most candidates treat as one. The first layer is the AI screening layer. Before a human reads your resume, it is in most competitive hiring processes evaluated by an applicant tracking system that uses AI to assess relevance, keyword alignment, and structural quality relative to the job description. Candidates who do not clear this layer do not get read. The optimisation required to clear it is specific, learnable, and almost entirely neglected by conventional resume advice.
The second layer is the human layer. Once a resume clears the AI screen, a human reads it, typically in six to eight seconds on first pass, making a binary decision about whether it warrants further attention. The optimisation required to perform well on this layer is different from the first layer and in some respects in tension with it. A resume optimised purely for keyword density can clear the AI screen and fail the human read. A resume written purely for human impact can fail the AI screen and never be read at all.
The prompts below address both layers in sequence. They are designed to be run for a specific job description rather than as a generic resume improvement exercise, because a resume optimised for a specific role consistently outperforms a strong general resume in competitive hiring processes by a margin that the research makes clear.
A 2024 analysis of 640,000 job applications by the recruitment analytics firm Jobscan found that applications with resume-to-job-description match scores above 75% were 3.1 times more likely to receive an interview invitation than applications with match scores below 50%, controlling for candidate experience level and role seniority. The match score is not a measure of qualification. It is a measure of how well the resume's language mirrors the language of the job description. Two equally qualified candidates with different match scores get different outcomes. The one who understands this and acts on it has a structural advantage that has nothing to do with their underlying capability.
Prompt 1: The job description decoder
The problem it solves: understanding what a job description is actually looking for beneath the surface language, including the unstated priorities that the explicit requirements do not fully capture.
You are a senior talent acquisition specialist
with extensive experience interpreting job
descriptions and understanding what hiring
managers actually prioritise when they
write them.
Here is the job description I am applying for:
[paste full job description]
Please analyse it as follows:
1. The explicit requirements: the qualifications,
experience, and skills stated directly,
ranked by how prominently they appear
in the description
2. The implicit requirements: capabilities
or characteristics the description
signals without stating directly,
based on the language used, the
problems described, and the context
of the role
3. The keywords and phrases that an
applicant tracking system is most
likely to be screening for,
listed in order of likely weighting
4. What this role is actually trying
to solve: the underlying business
problem or organisational need
that prompted this hire
5. The profile of the candidate most
likely to be shortlisted, described
in specific terms rather than
generic capability language
Tell me anything in this job description
that is unusual, that signals something
specific about the organisation's
current situation, or that I should
factor into how I position my application.
Point four, what the role is actually trying to solve, is the output that most candidates never develop and that most interviewers wish more candidates had thought about. A candidate who opens an interview by demonstrating that they understand the underlying business problem the role was created to address is having a different conversation than one who demonstrates they meet the stated requirements. Both conversations can lead to an offer. The first one more reliably does.
Prompt 2: The experience mapper
The problem it solves: identifying which elements of your professional history are most relevant to a specific role and how to frame them in the language the hiring process is optimised to recognise.
You are a professional resume writer helping
me identify and frame the most relevant
parts of my experience for a specific role.
The job description analysis from the
previous step: [paste output from Prompt 1]
My current resume or professional history:
[paste your current resume or a detailed
description of your experience]
Please:
1. Identify the three to five experiences
in my history that are most directly
relevant to what this role requires,
including experiences I may have
underweighted or described in language
that does not reflect their relevance
to this specific role
2. For each relevant experience, suggest
how it should be framed using the
language and priorities identified
in the job description analysis,
without changing the underlying facts
3. Identify any gaps between my experience
and the role requirements that are
significant enough to require a
proactive response, either in the
resume or in a cover letter
4. Identify any experiences I am currently
including in my resume that are
unlikely to add value for this specific
role and that are taking up space
that higher-relevance content could use
Constraint: do not suggest I claim
experience I do not have.
The goal is accurate reframing,
not misrepresentation.
The constraint against misrepresentation is not a formality. AI will suggest embellishments when asked to strengthen a resume without this explicit guardrail, because strengthening the resume and accurately representing the candidate's experience are not always the same objective. The constraint keeps the output honest and keeps the candidate out of a position they cannot sustain once in the role.
Prompt 3: The ATS optimiser
The problem it solves: restructuring resume content to maximise keyword alignment with a specific job description without producing the kind of keyword-stuffed text that fails the human read.
You are an applicant tracking system specialist
helping me optimise my resume for both
automated screening and human review.
The target keywords and phrases from
the job description: [paste from Prompt 1 output]
My current resume content for the
most relevant roles: [paste the experience
sections most relevant to this application]
Please rewrite each section to:
1. Incorporate the high-priority keywords
naturally within achievement statements
rather than in a skills list or
keyword section that reads as mechanical
2. Lead each bullet point with a strong
action verb that reflects the language
register of the job description
3. Quantify impact wherever the content
supports it, using conservative
estimates with appropriate qualifying
language where precise figures
are not available
4. Remove or consolidate any content
that does not contribute to the
relevance score for this specific role
Constraints: every statement must be
factually accurate. No bullet point
should exceed two lines. The rewritten
content should read naturally to a
human reviewer, not as a keyword list
dressed in sentence form.
The instruction to incorporate keywords within achievement statements rather than in a dedicated skills section reflects a specific and important finding about how modern ATS systems evaluate resumes. Earlier generation systems rewarded keyword density regardless of context. Current systems are increasingly capable of assessing keyword relevance in context, and a keyword appearing within a specific, quantified achievement statement is weighted more heavily than the same keyword in a generic list. The optimisation strategy has evolved and most generic resume advice has not kept up with it.
Prompt 4: The human read optimiser
The problem it solves: ensuring that the resume that clears the AI screen also performs well in the six to eight seconds a human reviewer spends on first pass.
You are a hiring manager reading this resume
for the first time. You have six seconds
to decide whether it warrants further attention.
Here is the resume as currently drafted:
[paste current draft after ATS optimisation]
The role you are hiring for:
[paste job title and two-sentence summary]
Please give me your honest six-second reaction:
1. What is the single most compelling
thing about this resume that you
noticed immediately
2. What is the first question or doubt
it raised in your mind
3. What is the most important thing
that is currently buried that
should be near the top
4. What would make you put this in
the yes pile without reading further
Then give me specific recommendations
for restructuring the resume's hierarchy
so that the most compelling content
appears where a six-second reader
will see it.
Be direct. A resume that does not
survive six seconds does not get read.
The six-second constraint is not a rhetorical device. Eye-tracking research on resume review behaviour, most recently a 2024 study by TheLadders using eye-tracking technology with 30 experienced recruiters, found that recruiters spent an average of 6.2 seconds on initial resume screening before making a proceed or discard decision. The content visible in those 6.2 seconds determined the outcome in 94% of cases. Everything below the fold on the first page is read only by candidates who have already cleared the initial screen based on what appeared above it.
Prompt 5: The cover letter connector
The problem it solves: writing a cover letter that does the specific job a cover letter is actually meant to do, which is not to summarise the resume but to make the case for why this specific candidate is right for this specific role in a way the resume cannot.
You are helping me write a cover letter
that makes a specific, compelling case
for my candidacy rather than restating
my resume in paragraph form.
The job description analysis:
[paste from Prompt 1]
The underlying business problem
this role is solving: [from Prompt 1]
My most relevant experience
and how it maps to the role:
[paste from Prompt 2]
Any gaps I need to address proactively:
[from Prompt 2]
Please write a cover letter that:
1. Opens with a direct statement of
why this specific role at this
specific organisation is relevant
to the work I do, not a generic
expression of interest
2. Makes one specific, substantiated
case for why my background addresses
the underlying business problem
the role is trying to solve
3. Addresses the most significant gap
between my experience and the
requirements directly and briefly,
without dwelling on it
4. Closes with a specific, confident
ask rather than a passive expression
of hope
Constraints: under 300 words.
No sentences that could appear
in any other cover letter for
any other role. No "I am writing
to express my interest in."
No "I believe I would be a great fit."
Every sentence must earn its place
by saying something that is both
true and specific to this application.
The constraint that no sentence should be capable of appearing in any other cover letter for any other role is the most demanding and most valuable constraint in this prompt. It is also the constraint that most AI-generated cover letters violate, because generality is the default mode of language models when producing professional documents. Enforcing specificity requires explicit instruction, and the instruction needs to be as clear as this one to produce output that is genuinely differentiated.
The ten-minute schedule
Prompt 1: three minutes, including reading the output carefully enough to use it as input for subsequent prompts. Prompt 2: two minutes. Prompt 3: two minutes per role section, applied only to the two or three most relevant roles. Prompt 4: one minute, applied to the full draft. Prompt 5: two minutes, including editing the output for your specific voice.
The ten minutes assumes you have a current resume to work from. The first time you run this sequence for a new application, budget fifteen minutes. By the third application, ten minutes is consistently achievable.
What the sequence produces is not a perfect resume. It is a resume that is significantly better calibrated to the specific role than a generic strong resume would be, in a time investment that is small relative to the stakes of the application.
The resume is not the job. It is the door. These prompts help you get through it.
Monday we are returning to a theme that has run beneath the surface of this newsletter since Issue #1 and that the past several issues have been building toward: the specific moment in a career when the accumulation of AI capability, constraint-trained problem-solving, and strategic visibility converges into something that looks, from the outside, like an overnight success.
It is not overnight. It is the compound interest on decisions made earlier, when the return was not yet visible. Monday's issue is about recognising where you are in that compounding curve and what the next decision should be.
— The Artificial Idea team
