🚀 The pace of technological change is no longer linear, it's exponential.

From automation to Agentic AI, we're witnessing a transformative shift across industries — reshaping how we work, hire, and scale.

As part of this shift, I recently built an AI-powered recruitment screening workflow using n8n — one that could meaningfully support the Talent Acquisition function, not replace the judgment it depends on.

❓ PROBLEM Most companies still handle recruitment manually — screening CVs one by one, replying to applicants, checking for duplicates, and manually tracking which positions are still open. It's repetitive, slow, and error-prone — and it eats hours of HR's time every week.

💡 SOLUTION I built an AI-powered recruitment automation that: ✅ Automatically reads incoming applications from Gmail ✅ Matches and validates the position applied for against open roles — even with typos or inconsistent naming ✅ Extracts and evaluates each CV against HR-defined criteria using AI ✅ Detects duplicate applications and responds accordingly ✅ Informs applicants when a position is closed, and shares current openings ✅ Routes candidates into three tiers — Recommended, Consider, Reject ✅ Logs every application and outcome for full traceability ⚡️ RESULT (based on real testing across 8 scenarios) ✅ Processes a full application — reading the email, matching the position, extracting the CV, running the AI evaluation, and replying — in an average of ~25 seconds ✅ Early rejections like duplicates and invalid applications resolve in under 15 seconds ✅ Even non-standard CVs (image-based/creative formats) still get processed automatically, falling back to a secondary extraction method — takes ~40 seconds, still well under a minute ✅ Borderline ("Consider") candidates reach HR's review queue in roughly the same ~25-30 seconds — the only wait left is a human decision, not the system ✅ Every application, across all 8 tested scenarios, is logged and traceable — nothing gets lost or forgotten ✅ Runs continuously with no dependency on someone being available to trigger it

🧠 WHY IT'S DIFFERENT This isn't a "trust the AI blindly" system. AI can misread or hallucinate — so every ambiguous case gets flagged and routed to a human instead of being silently auto-rejected. The goal is removing repetitive work, not removing human judgment where it actually matters.

🔧 WHAT'S NEXT I'm continuing to refine the evaluation logic based on real test cases, and exploring how to extend this into other parts of the hiring pipeline.

💡 AI isn't just a tool that speeds up existing processes — used well, it becomes part of how a workflow makes decisions. The real work is designing where AI should lead, and where a human still needs to.

🤝 If you're exploring how n8n or AI-powered workflows could fit into your hiring process (or anywhere else in your ops), feel free to drop me a message — happy to connect and compare notes.

Check it out