How AI Suggestions and Job Matching Work
Cvisioner offers three AI capabilities: content suggestions, ATS analysis and job matching. This article explains how each works and how your data flows.
Content suggestions
When you write a sentence in the CV editor, the AI proposes alternatives in the side panel. Suggestions are personalised to your target role and your existing experience.
The suggestions aim to:
- use action verbs (for example "built", "designed", "led" instead of "did")
- add measurable outcomes (for example "30% conversion increase", "team of 5")
- bring ATS-friendly keywords forward
You apply a suggestion with one click or keep writing your own sentence. The AI does not decide for you — it only offers alternatives.
ATS scoring
The "ATS Score" tab rates your CV from 0 to 100 for how well applicant tracking systems (Workday, Greenhouse, Lever and similar) can read it.
The score is built from four axes:
- Keywords: variety of skills, technologies and certificates
- Formatting: structural completeness and parseability (contact details, sections)
- Experience: content quality per bullet — action verbs, measurable results, density
- Skills: depth and breadth (skills, languages, certificates, projects)
Score bands match what the app shows: 55 and above is "good", 80 and above is "strong"; below that, "needs work".
ATS analysis is a Pro feature.
Job matching
The "Job Matching" tab compares the skills, experience and education in your CV with open job postings.
The match is computed in two layers:
First layer — keyword match: technical skills in your CV (React, Python, AWS and so on) are compared one to one with the requirements in the posting.
Second layer — semantic similarity: sentences are compared by meaning. A posting looking for a "backend developer" also recognises your "server-side developer" experience.
The weighted average of the two layers is shown as a score from 0 to 100. 55 and above is marked as a "strong match" (green ring), 38 to 54 as "moderate" (amber); strong matches come first in the list. The score is visible for every job on every plan.
Data flow and privacy
Your CV content is sent to the Google Gemini API for analysis. Google does not use this data for model training — this is contractually bound by a Data Processing Addendum (DPA).
Data flow:
- You write content in the CV editor
- The content is encrypted and stored in Supabase (EU-Frankfurt)
- When you trigger an AI feature, the content is sent temporarily to the Gemini API
- The response is shown in the panel; it is not stored permanently at Gemini
The details of this flow are defined in the KVKK Privacy Notice and the Explicit Consent text. You can withdraw your consent to AI processing at any time; AI features are then disabled while basic CV building keeps working.
Which AI feature is on which plan
There are no counters or daily quotas; the difference is per feature.
Free plan: a match score and reasoning is visible for every job; the details and fit analysis of the 4th and 5th jobs in your ranking are open (or, if neither fits your profile, jobs in your ranking that do fit are opened first) and a job-specific CV and cover letter are prepared on screen for those two; download is on Pro. ATS analysis, gap analysis and the AI writing tools in the CV editor are not part of the free plan; touching them opens the upgrade screen.
Pro plan: details for all matches, Ready Applications, ATS analysis, gap analysis, AI writing tools (summary, experience bullets, skill suggestions) and a cover letter written for the job, CV Studio (creating new CVs, templates, PDF and Word download), Job Analysis.
During onboarding (until the profile wizard is completed) AI features are open regardless of plan.