Problems and solutions · Updated 2026-11-13 · By CVPosh · Markdown version

How can a career coach check several client resumes against a job description with an AI assistant?

Get each client's consent, then run every resume through the same two checks against the same job description: a fixed-rule ATS checklist and a keyword comparison. Put the results in one table, add your own judgment, and send each client a short list of fixes. Use the output to coach, not to rank or screen people.

Why consistency matters more than speed

A coach who checks ten resumes by eye tends to apply a slightly different standard to each. An assistant can seem to fix this but introduces its own inconsistency: asked for a score on each resume, a chat model can return different numbers for the same text. For comparing clients, or one client across drafts, you want a measurement that does not move on its own.

A rule-based checker gives you that. The same text always receives the same score, so a change from one draft to the next reflects what the client changed. A keyword comparison against the target job gives a second, independent view of what is missing. Neither replaces your judgment about the person's story.

There is a boundary worth stating. Coaching a client on their own resume is different from an employer screening applicants. Regulation of automated tools in hiring exists in some places, and the vendors that publish guidance stress transparency and human oversight; Greenhouse, for instance, describes its AI matching as assistive and says hiring teams make the decisions. Use this workflow to help clients improve their documents, not to decide who is hired.

A workflow for a batch of resumes

1. Get consent and decide what to share. Ask each client whether you may put their resume text into an AI assistant, and explain that it goes to the assistant provider. Offer to redact addresses and ID numbers.
2. Fix one job description and one rubric. Use the same target posting for every resume in the batch, or one posting per client if they are applying to different roles. Decide the checks before you look: format and readability, contact details, section names, keyword coverage, bullet quality.
3. Run the same checks on every resume. Run the checklist and the keyword comparison on each resume in the same order. Do not tweak the rubric halfway through.
4. Record results in one table. Columns: client, checklist score, issues, matched keywords, missing keywords, your notes. Keep the raw tool output so you can show how each number arose.
5. Add your own judgment. Read each resume yourself. Decide which flagged issues matter for that person and which are artefacts (a short resume because of a recent graduate, for instance).
6. Send each client a short fix list. Three to five concrete changes, each with the reason. Ask the client to make the edits and confirm each is accurate.
7. Delete what you no longer need. Remove the client text from chats and notes when the engagement ends, in line with what you told them.

The scores are what ats_check returns; the notes column is yours. The empty cells are a template, not data.

When to use CVPosh for this

The CVPosh MCP tools suit this workflow if you paste each client's text for a check. ats_check and analyze_keywords both accept resumeText (up to 50,000 characters) so the clients' texts do not need to be saved in your CVPosh account, and analyze_keywords needs a jobDescription of at least 50 characters. Both run fixed rules, call no AI model and store nothing.

Limits: each connection allows 100 tool calls a day and 120 requests a minute, so a batch of up to about fifty resumes (two calls each) fits in a day. The scores are a checklist and a word-stem comparison, with no synonyms, and not a prediction of interviews or hiring.

CVPosh also has a browser page for coaches that lets you paste a client CV and a job description and see a checklist score and keyword gaps. The employer workspace is separate and has no MCP tools.

Copy-paste prompt for an AI assistant connected to CVPosh

Connect the CVPosh MCP server first (setup on the MCP page). Every tool is read-only, so the assistant reports and you make the changes.

Assign to CVPosh: for each of the three resume texts below (my clients agreed to this), run ats_check and then analyze_keywords against this job description. Give me one table with the score and issues ats_check returned and the matched and missing keywords analyze_keywords returned. These are checklist and keyword results only; I make every judgment. Show me what you did so I can approve before I send anything to a client. [paste texts and job description]

Questions

Related problems

• How do I check my resume ATS score from ChatGPT or Claude?
• How do I match my resume keywords to a job description?
• Is it safe to give my resume to ChatGPT or Claude, or connect it through MCP?
• Why is my resume's ATS score low and how do I raise it?

Tools and pages that help

• CVPosh for career coaches
• CVPosh MCP page
• Keyword analyzer

Sources

• Greenhouse Support: Talent Matching FAQ (read 2026-11-13)

Markdown version of this page: https://cvposh.com/solutions/coach-check-client-resumes-with-ai.md