← Вернуться на сайт

Ops · HR · Automation

HR Automation: How to Process Resumes and Track Time with Scripts

How to use LLMs and Google Scripts to automate HR routines without programming skills.

Why It Works

Google Workspace has long been a hidden 'corporate automator.' Sheets, Drive, Forms, Gmail, Calendar — all are connected into one ecosystem via Google Apps Script.

The problem used to be one thing: scripts had to be written manually. An HR manager, recruiter, or operations manager usually doesn't do that.

Now, LLMs (Gemini, ChatGPT, Claude) remove this barrier:

  • you explain the task in simple words;
  • the model generates a working script;
  • you paste it into Google Scripts;
  • an automation starts that previously seemed 'only for programmers.'

In effect, LLMs turn Apps Script into 'voice-activated automation.'

Where HR Spends the Most Time

  1. Downloading and sorting resumes.
  2. Entering data into spreadsheets.
  3. Manual triage of candidates.
  4. Reminders, emails, follow-ups.
  5. Tracking employee time.

All these tasks are repetitive. And what is repetitive = can be automated.

How "15-Minute" Automations Work

The mechanics are simple:

  1. Describe the task to the LLM like this: → 'When a resume from hh.ru is dropped into a Google Drive folder, you need to: — identify the job opening; — place the file in the corresponding folder; — extract key data; — write a row to Google Sheets.'
  2. The model generates a script.
  3. You paste it into Apps Script.
  4. The script runs regularly (on a schedule or event).
  5. HR gets structured data, not PDF files.

Critically important point: HR doesn't write code. They describe a process. The model does the rest.

Example 1: Automatic Resume Sorting

Scenario

Standard resumes from hh.ru have a similar structure: Full Name → Position → Experience → Contacts.

What the Automation Does

  • The 'Incoming resumes' folder becomes the input;
  • An LLM prompt generates a script that:
    • identifies the job opening by keywords;
    • moves the resume to the correct folder → “Designer”, “PM”, “Marketing”;
    • records the fact in a spreadsheet: name, email, experience, date received.

What HR Gets

Not ten PDFs in an email, but: — a structure by job openings; — a table with candidates; — a signal that there are new resumes.

Example 2: Automatic Candidate Triage

Triage is a logical check:

  • are the necessary skills present;
  • is the experience suitable;
  • which stage to send them to.

How LLM Helps

You describe the rules: → 'If Java, Spring, Kubernetes are listed — Senior. If HTML/CSS, one year of experience — Junior. If nothing matches — Reject.'

The LLM turns this into a script that:

  • reads the spreadsheet;
  • sets statuses;
  • marks 'suitable,' 'questionable,' 'rejected.'

HR Only Does One Thing

Looks at the finished table and writes to the suitable candidates.

Example 3: Notifications in Slack or Gmail

When a new candidate appears (or a status is updated), HR needs to know 'now,' not half a day later.

The LLM generates a script that:

  • tracks new rows;
  • sends a short notification to HR in Slack or email;
  • includes the name, job opening, and a link to the file.

The principle is one: 'less manual scrolling — more decisions.'

Example 4: Simple Time-Tracking via Forms + Sheets

The Problem

Time tracking often turns into chaos: Excel files, messengers, voice messages.

The Solution

  1. Google Form: 'what did you do,' 'how many hours,' 'project.'
  2. Responses automatically fall into a spreadsheet.
  3. The LLM creates a script that:
    • calculates hours for the week;
    • aggregates by project;
    • creates a summary on a separate sheet;
    • sends a report to HR every Monday.

This replaces expensive tracking systems.

Example 5: Automatic Emails to Candidates

The most tedious part is follow-ups:

  • invite;
  • reject;
  • clarify;
  • send a test assignment.

The LLM generates a script that:

  • takes the email from the spreadsheet;
  • substitutes the email template;
  • sends the email based on a trigger (e.g., when the status is 'Interview').

You write the email once → the script does the work from then on.

Why LLMs Lower the Barrier to Entry for Automation

1. No Need to Know JavaScript

You describe the business logic — the model writes the code.

2. Easy to Explain Changes

'Add a filter for the Python skill' — the model will rewrite the necessary piece.

3. Scripts Become 'Live'

They can be tweaked just like talking to a person.

4. Fast Iterations

Any edit → a couple of minutes.

5. HR Stops Being a 'Manual Processor'

And starts dealing with people, not data entry.

Typical Processes That Can Be Automated in a Day

  • resume sorting;
  • skill flags (Python, Figma, English);
  • candidate stage checklist;
  • sending test assignments;
  • status reports;
  • reminders for interviewers;
  • employee time summary;
  • new application notifications;
  • auto-filling candidate cards;
  • duplicate cleanup.

If a process is repetitive, it can be automated.

Conclusion

Google Workspace + LLMs turn HR routines into a set of buttons:

  • describe the process →
  • get the script →
  • launch the automation.

Without programming skills. Without an IT department. In 15 minutes.

The main value: HR gets their time back to work with people, not with files.