Free Self-Evaluation Phrase Catalog | 200+ Phrases by Skill, Role & Rating
Free Self-Evaluation Phrase Catalog | 200+ Phrases by Skill, Role & Rating
To use AI to find a job, let it do the research, matching and drafting, and keep the decisions and the sending for yourself: define your target role, let AI collect and score open roles, tailor your CV to each one you actually want, and track the replies. A free career agent such as Atlas Career covers the first three steps at no cost. What does not work is mass auto-applying: the average posting on Greenhouse drew 228 applications in February 2024 (Axios), and generic ones are filtered out first.
This guide gives you the 5-step plan, a 7-day starter schedule, a worked time budget and the mistakes recruiters complain about. If you are tempted by speed, read the AI auto-apply risks first.
The 5-step plan: how to use AI to find a job
| Step | What AI does | What you do | Output |
|---|---|---|---|
| 1. Define the target | Turns your experience into role options and a skills list | Decide must-haves: role, level, place, salary, no-gos | One-page target profile |
| 2. Find and score roles | Reads many postings, spots patterns, ranks fit | Pick 5 to 10 roles you would accept if offered | Shortlist |
| 3. Tailor the application | Drafts CV bullets and a cover letter from your real history | Check every fact, add one company-specific line | Application per role |
| 4. Reach people | Drafts short messages to recruiters and team members | Personalise and send a few, not fifty | Conversations |
| 5. Track and learn | Summarises status, drafts follow-ups, spots patterns | Approve every message, follow up once per stage | Weekly review |
Step 1: Define the target before you open a chatbot
AI amplifies whatever focus you bring. Without one, you compete on speed against hundreds of others. Start with a conversation, not a search:
- Paste your CV (without ID numbers or employer secrets) and ask: “Name my 10 strongest skills and 5 achievements, with the evidence from the text.”
- Ask: “Suggest 5 realistic job titles for this profile in the US and in Europe, and what each one requires.”
- Write down must-haves: role, seniority, location or remote, salary band, language, visa. If a role is a “maybe”, it stays off the list.
- Keep one “source of truth” document with your real history, so every AI draft stays consistent with it.
If you want to see where your LinkedIn profile stands first, use the guide on AI for LinkedIn profiles.
Step 2: Find and score roles with AI
There are three ways to get roles in front of you, and they combine well:
| Way to find roles | Strength | Limit | Cost |
|---|---|---|---|
| Chat assistant (ChatGPT, Claude, Gemini) | Analyses postings you paste in; finds patterns and keywords | Does not search live jobs reliably; can invent details | Free to paid |
| AI matching on job boards | Shows roles that fit your filters; official portals such as the Bundesagentur für Arbeit are free | Match logic is rarely explained | Free |
| Career agent (Atlas Career) | Searches 1.5M+ active jobs straight from employers' career sites, scores each with reasons | You still decide and apply yourself | Free, no limit |
| Auto-apply bots | Sends applications in volume | Generic, error-prone, often read as spam | Usually paid |
Practical routine: collect 5 to 10 postings for one role and region, then ask a chat assistant: “Separate must-haves from nice-to-haves, and tell me which keywords appear in at least 4 of 5 ads.” Example: for five entry-level data engineer ads, the answer might show Python, SQL and AWS in four of five, plus German language in the Germany-based ones. You then update your CV only where it is true. Then validate with real checks: the company's career page, employee posts and, where possible, a direct conversation.
To see what a career agent does in practice, open the Atlas job pages: roles sorted into four portals (healthcare, trades, sales, tech and product), each with a newest-first list.
Step 3: Tailor each application, then edit hard
This is where smart help turns into obvious spam. Use AI for structure and phrasing, never for truth, tone or the final decision. Recruiters do not mind AI use; they mind generic, careless output. In a Robert Half survey of March 2026, 67% of HR leaders said AI-generated applications are slowing hiring.
- Paste CV and job ad and ask: “Which 6 to 8 bullets match this role best? Do not invent anything.”
- Keep the layout simple and ATS-friendly: one column, standard headings, no text in images.
- Draft a cover letter of 3 to 4 short paragraphs tied to the 3 main requirements.
- Add one concrete detail about the company and delete generic filler.
- Read it aloud. If it does not sound like you, rewrite it.
Career switcher example: a marketer moving into analytics asks AI to emphasise testing and data storytelling from real projects, then corrects two overstated bullets by hand. The result is a CV that is honest and targeted. For a longer do and don't list, see AI job application mistakes.
Step 4: Reach people, a few at a time
Many hires start with a person, not a form. Use AI to draft, then rewrite: a two-sentence note that names the role, one relevant result and one question. Send five personal messages a week instead of fifty templates. Never let a tool send messages or follow-ups on its own.
Step 5: Track and follow up like a human
Use a spreadsheet or a tracker with company, role, link, date, version sent, status, contact and next action. Once a week, ask AI to prioritise follow-ups and draft short skeletons by status; you edit and send. Follow up once per stage after 7 to 10 days, not daily. More patterns in the guide to AI job application trackers.
A time budget that works: worked example
Assumption (ours, not a measurement): you have 10 hours a week and need about 30 to 40 minutes per properly tailored application, including fact checks.
| Block | Hours / week | What it buys |
|---|---|---|
| Research, shortlist, tracking | 3 | 10 to 15 roles reviewed, 6 to 10 selected |
| Tailored applications | 5 | 6 to 10 applications at 30 to 40 minutes each |
| People and follow-ups | 2 | 5 personal messages, follow-ups due that week |
The same 10 hours in a bot would produce hundreds of applications, but each one would be one of roughly 228 on a posting. If after four weeks your interviews do not rise, tighten the targeting before you add volume.
Weekly starter plan: your first 7 days
- Day 1: target profile, must-haves, “source of truth” document.
- Day 2: collect 10 postings, extract must-haves and keywords, update CV where true.
- Day 3: create your career profile in a free agent such as Atlas, review the first matches.
- Day 4 and 5: one tailored application per day for roles with a high match.
- Day 6: five personal messages to people at target companies.
- Day 7: tracker clean-up, one thing that worked, one to change.
Weekly goals by situation: new graduates aim for 5 to 10 tailored applications and one screening call a week; experienced specialists for 4 to 8 and one to two calls; career switchers for 3 to 6 and one call every one to two weeks.
Atlas Career: a free option for the whole loop
Atlas is a free AI career agent built by Sprad in Vienna. You create your profile in a short voice conversation (role, level, salary, location, no-gos). Atlas then keeps searching over 1.5 million active jobs straight from employers, gives each role a match score from 0 to 100 with reasons, and tailors a CV, and a cover letter on request, for every application. It is free with no limits and sends nothing itself: you review each document, download it and apply directly with the employer.
Fair comparison: Atlas is not the only way. A chat assistant is fine for analysing a handful of postings, and job boards remain useful for local roles. What Atlas adds is search and tailoring in one place without a paywall, with data hosted in the EU. Data is never sold or used to train other AI models, and you can delete it at any time.
What to avoid: recruiter red flags
- Mass auto-apply to hundreds of roles: it signals “no fit filter”.
- The same cover letter everywhere: recruiters spot a template in seconds.
- Placeholder text or the wrong company name left in.
- Invented metrics, projects or scope: a quick route to rejection for integrity.
- Keywords you cannot back up: screening answers will not match your CV.
- US-casual tone for DACH roles where formal, factual German is expected.
- Automated bulk follow-ups that humans and systems flag as spam.
- AI answers generated live in interviews: this can be seen as cheating.
Watch out for fake offers. Honest employers never ask you to pay to get a job, and a check that you must partly send back is a warning sign. The US Federal Trade Commission lists the typical patterns (article of March 2023).
Special notes for Europe and DACH
- Data minimisation: redact ID numbers, client names and confidential figures before pasting anything into AI. The principle is in Art. 5(1)(c) GDPR.
- Employers use AI too: software for selecting applicants counts as high-risk under Annex III of the EU AI Act, so ask how your application is assessed.
- Tone and format: German applications often expect “Sie”, precision and a photo; US ones do not. Keep one template per market. See AI for job applications in Europe.
- Save what you sent: keep the exact files per application, so documents stay consistent across longer processes.
For tool picks, see the comparison of the best AI job search tools in 2026 and the Europe playbook.
Frequently Asked Questions (FAQ)
How can I use AI to find a job?
Follow five steps: define your target, let AI find and score roles, tailor each application, contact a few people, and track the outcome. AI does the research and the drafting; you choose the roles, verify the facts and press send.
Can AI find jobs for me, or only write applications?
Both, depending on the tool. Chat assistants help you analyse postings you give them. Career agents such as Atlas search live jobs from employers' career sites and score them. Either way, you should check the role and apply yourself.
Is there a free way to use AI for the job search?
Yes. Chat assistants have free tiers, official job portals are free, and Atlas Career is free without a limit. Check what happens after the first few tailored CVs, because many tools cap their free plan.
Are auto-apply bots a good idea?
For most professional roles, no. They produce generic documents at high volume, and recruiters filter them first. Use AI to speed up research and drafting instead, and apply to fewer roles with care.
How do I stop my AI-assisted applications from looking like spam?
Apply to fewer roles and tailor every one. Mirror keywords only where they are true, add one company-specific detail and remove filler. Proofread names, titles and dates before sending.
Is it safe to paste my CV into ChatGPT?
Remove ID numbers, addresses and confidential employer details first. In the EU, data minimisation applies, so prefer tools with clear EU hosting and an explicit statement about training use.
How many applications per week should I send?
For most people, 4 to 10 tailored applications a week is realistic with about 10 hours of effort. If interviews do not increase after four weeks, improve targeting before you increase volume.
Which roles benefit most from AI in the job search?
Any role where clear writing and evidence count: business, tech, operations, marketing and customer success. The benefit depends less on the role than on honest inputs, tight targeting and careful editing.
Jürgen Ulbrich
Jürgen Ulbrich has more than a decade of experience in developing and leading high-performing teams and companies. As an expert in employee referral programs as well as feedback and performance processes, Jürgen has helped over 100 organizations optimize their talent acquisition and development strategies.
