Post 10 · AI & Automation

How to Build an AI Resume Screening Workflow Using n8n, Google Sheets, and OpenAI

A step-by-step guide for recruiters to understand how AI-assisted resume screening works — and how automation can support faster, smarter shortlisting.

Recruiters spend hours reading resumes, comparing them against job descriptions, creating shortlist notes, and updating candidate trackers. It is one of the most time-consuming parts of the job — and it is often the part that delays the entire hiring process.

What if a significant part of that work could be automated?

This post walks you through how an AI-assisted resume screening workflow works — what tools are involved, what each step does, and what output it generates for the recruiter. At the end, you will understand exactly what AI can and cannot do in the screening process — and how to use it as a productivity tool without replacing recruiter judgment.

Important Disclaimer

AI resume screening should be used as a recruiter support tool, not as a final hiring decision-maker. Human review is still required to validate skills, experience, communication, context, and fairness. Never automate the final hiring decision.

What Is an AI Resume Screening Workflow?

An AI resume screening workflow is an automation pipeline that takes a candidate's resume text, compares it against a job description, and generates a structured screening output — including a match score, matched skills, missing skills, a fit assessment, and a recruiter summary — written straight back into your candidate tracker.

Instead of a recruiter manually reading 50 resumes and taking notes on each one, the workflow does the initial pass automatically. The recruiter then reviews the AI-generated summaries and makes the final shortlisting decision.

Think of it as having a very fast, very organised junior recruiter who never gets tired — but whose work you always double-check before acting on it.

Tools Required

⚙️
n8n
Workflow automation tool. Free and open-source. Connects all the tools together.
📊
Google Sheets
Your candidate tracker. You enter the candidate details; n8n reads new rows and writes the AI results back into the same row.
🤖
OpenAI API
The AI engine. Analyses the resume against the JD and generates structured output.
This is the simplest possible version of the workflow — everything lives in one Google Sheet. We will cover the Google Drive version (auto-reading uploaded resume files) in a later post. For now, the resume text is entered directly into the tracker.

Workflow Overview

Here is the complete flow, from a new candidate row to a finished screening inside your tracker:

The AI resume screening workflow built in n8n, showing the trigger, OpenAI screening node, output formatting, and Google Sheets candidate tracker update
The core workflow built visually in n8n — a new candidate triggers the AI screening node, the output is formatted, and the results are written back to the Google Sheets candidate tracker.
1
New candidate added to the tracker
You add a new row to the sheet with the candidate's details and resume text in the input columns
2
n8n detects the new row
The Google Sheets trigger node fires automatically whenever a new row is added (Row Added)
3
Resume Text + Job Description sent to OpenAI
The AI Resume Screening node sends both fields to OpenAI with a structured prompt asking it to compare them
4
Output is formatted into clean fields
The Format Screening Output node parses the AI's response into a match score, fit category, skills, summary, and recommendation
5
Results written back to the same row
The Update Candidate Tracker node writes the AI output columns back into the candidate's row — no manual data entry

Step-by-Step Setup

Step 1 — Create the Tracker Google Sheet

Create a Google Sheet and name it AI Resume Screening Tracker. It has two groups of columns: the input columns you fill in for each candidate, and the AI output columns that n8n writes back automatically. Set them up in this order, left to right:

Input columns — you fill these in

ColumnWhat It Stores
Candidate NameThe candidate's full name
EmailCandidate's email address
PhoneContact number (format as plain text so it isn't converted to a number)
LinkedIn URLLink to the candidate's LinkedIn profile
Job TitleThe role being screened for
Job DescriptionThe full JD the resume is compared against
Resume TextThe candidate's resume pasted in as plain text

AI output columns — n8n fills these in

ColumnWhat It Stores
AI Match Score0–100 score generated by the AI
Fit CategoryStrong Fit / Good Fit / Average / Not Suitable
Matching SkillsSkills on the resume that match the JD
Missing SkillsSkills in the JD not found on the resume
Recruiter SummaryA short plain-English summary for the recruiter
RecommendationSubmit / Screen Further / Reject
StatusMarks the row as AI Screened once processing is done

Step 2 — Add a Candidate

Fill in the seven input columns for a candidate — name, email, phone, LinkedIn, job title, the job description, and the resume text. Leave the AI output columns blank; the workflow fills those in. Adding the row is what kicks everything off.

Step 3 — Set Up the Google Sheets Trigger in n8n

In n8n, add a Google Sheets Trigger node. Connect it to your Google account, select the AI Resume Screening Tracker sheet, and set it to trigger on Row Added. Now every new candidate row runs the workflow automatically.

Step 4 — Send Resume Text + Job Description to OpenAI

Add an OpenAI node (this is the AI Resume Screening node in the diagram above). Use a Chat model — GPT-4o is recommended for accuracy. Feed it the Resume Text and Job Description from the new row, with a structured prompt that asks it to compare the two and return a match score, fit category, matching skills, missing skills, a recruiter summary, and a recommendation. The quality of your prompt directly determines the quality of your screening output.

Step 5 — Format the Screening Output

Add a Code or Edit Fields node (the Format Screening Output node) to parse the AI's response and map each value to the matching tracker column — score to AI Match Score, category to Fit Category, and so on.

Step 6 — Update the Candidate Tracker

Add a Google Sheets node set to Update Row (the Update Candidate Tracker node). Point it at the same row that triggered the workflow and write the AI output columns back in, then set Status to AI Screened. The recruiter opens the sheet and sees a fully screened candidate — no manual data entry.

What the AI Output Looks Like

Here is an example of what the AI generates for a Java Developer candidate already in the tracker. Each field maps directly to one of the AI output columns:

{
  "candidate_name": "Rahul Sharma",
  "ai_match_score": 82,
  "fit_category": "Good Fit",
  "matching_skills": ["Java", "Spring Boot", "REST APIs", "Microservices", "SQL", "Git", "Agile"],
  "missing_skills": ["AWS depth / production experience", "scalable cloud backend architecture"],
  "recruiter_summary": "Solid Java backend candidate with 6 years of experience and strong alignment to the core stack. Domain experience in banking and e-commerce. Main gap is that AWS exposure looks basic rather than strong.",
  "recommendation": "Submit",
  "status": "AI Screened"
}

n8n writes these values straight into the candidate's row. The recruiter opens the tracker, reviews the score and summary, and decides whether to call the candidate. The AI did the first pass — the recruiter makes the final call.

What Recruiters Can Use This For

Limitations to Be Aware Of

Know These Before You Rely on the Output
  • AI may miss context — a candidate's project experience may not be captured fully in a text-based resume
  • Resume text has to be added to the sheet — this version reads the Resume Text column, so the text must be pasted in first (a later Google Drive version can automate that step)
  • Output quality depends on the prompt — a poorly written prompt produces unreliable screening results
  • AI cannot assess communication or culture fit — those require a human conversation
  • Never automate the final decision — AI assists the recruiter. The recruiter decides.

Why This Matters for the Future Recruiter

The future recruiter will not be replaced by AI. The future recruiter will be replaced by another recruiter who uses AI better.

The skills that will matter in recruitment over the next 5 years are not just Boolean search and sourcing. They include:

At StaffIQ, we are building training programmes that prepare recruiters for exactly this — not just the technical skills of today, but the AI-assisted workflow skills of tomorrow.

Coming Soon

AI Resume Screening Workflow — Hands-On Workshop

Build this complete workflow from scratch — live, step by step, with real resumes and real job descriptions. StaffIQ is planning a practical hands-on workshop for recruiters and HR learners.

💬 Express Interest on WhatsApp 📅 Book a 1:1 Call

Get Posts in Your Inbox

Free weekly newsletter — practical recruitment tips every week.