HR-агенты против людей: почему ваш следующий оффер может прислать бот

Remember that feeling when you get another LinkedIn message that looks like it was written by a bot? It probably was. But if these used to be primitive scripts, today full-fledged AI agents are taking the stage. They don't just send spam—they analyze your code on GitHub, read your articles, and try to understand how well you'll fit the culture of a company you only know from the job description. The irony is that recruiters who spent years automating their work to spend less time on people ultimately created tools capable of leaving them out of work.

Let's figure out what exactly changed. We've moved from the "smart search" stage to the "autonomous action" stage. Modern AI agents for recruitment aren't just a layer on top of a database. They're systems that can independently set themselves tasks: find ten Python developers with fintech experience, conduct initial screening via chat, evaluate the technical assignment, and schedule a meeting with the team lead. The human becomes a redundant link in the most labor-intensive stages. If recruiters used to spend hours reviewing hundreds of resumes, now an agent does it in seconds—and it doesn't get tired or develop tunnel vision by Friday evening.

Why is this happening right now? The technology stack has matured. The emergence of large language models has allowed machines to understand the context of human language. If searches used to work by keywords, now an agent understands that "developing high-load systems" and "scaling the backend for millions of users" are roughly the same thing. This kills the sourcer profession in its classic form. Companies see massive savings in this: why maintain a staff of ten HR specialists if one experienced manager and a couple of AI service subscriptions can handle the same hiring volume?

But here lies the main trap. AI agents inherit all the biases present in the training data. If an algorithm sees that in the past a company more often hired graduates from certain universities, it will ruthlessly filter out everyone else, even if they're self-taught geniuses. We risk creating a system where a "non-standard" candidate never breaks through the digital filter simply because they don't match the statistical norm. The problem is that recruitment is always a bit about intuition and "chemistry," which we still can't digitize.

The rules of the game are also changing for candidates themselves. Now you need to optimize your resume not just for keywords, but for LLM logic. We're entering a strange era where one AI writes a resume for a candidate, and another AI checks it. In this battle of algorithms, human communication becomes a luxury product. Most likely, live recruiters will remain only in the Executive Search segment, where you need to convince top executives over dinner, not just send a link to a test assignment in Telegram.

Ultimately, AI agents aren't the end of HR but its harsh transformation. Those who are used to working as "file couriers" will have to find new work. And those who know how to build relationships and understand human motivation will have to learn to manage a swarm of bot helpers. Technology once again reminds us that when working with people, the most important thing is still people, not processes.

The main point: AI agents will completely take over the routine work of sourcing and screening within the next two years. Will there still be room for human empathy in this process, or will we finally turn into lines in a database?