Why Your Recruiting AI Is a Legal Minefield and a Money Pit (And What to Do About It)
90% of US companies missed their hiring goals in 2025. The average cost to hire? Nearly $4,700 per person. Some roles cost three to four times that. And meanwhile Workday faces a nationwide class-action lawsuit over alleged age discrimination in its AI hiring tools. Amazon scrapped its own AI recruiting tool in 2018 after it learned to discriminate against women. This isn't the future. This is the present and it's a mess.
Your AI Recruiter Isn't Smarter. It's Just Dumber and Riskier
Companies think AI will fix hiring. It often makes it worse. Amazon's AI recruiter looked at a decade of historical data. Because most applicants were men, the model learned to downweight resumes that included words like women's colleges. It learned gender bias from training data and amplified it. A decade later we're still seeing the same pattern with Workday's AI tools. The lawsuit alleges age discrimination. The model didn't mean to be racist or ageist. It just learned patterns that correlated with protected classes. This is the problem with AI hiring tools today. They're opaque black boxes that make decisions without explaining themselves. When you fire someone or reject a candidate, you can justify it. When an AI system rejects thousands of people, you can't. You're just betting that the math works out. It usually doesn't.
The Hidden Costs Are Crushing Your Bottom Line
Slow hiring costs companies $98 a day per open role. The average direct cost per hire is around $4,100. That's just direct recruiting costs. It doesn't count the time managers spend interviewing, the lost productivity while roles sit empty, or the chaos when teams can't scale. Mid-sized companies waste over 77,000 hours a year on manual HR tasks. That's 88 full-time employees doing nothing but data entry. Manual data entry costs U.S. companies $28,500 per employee annually. That number is shocking. It means the average employee spends weeks a year just typing data into spreadsheets and systems. This is the part nobody talks about. HR teams are drowning in paperwork while the real work, connecting with humans, is getting buried under administrative debt. The result is longer time-to-hire, worse candidate experiences, and more bad hires. Companies that think they're saving money by automating are usually just automating their own inefficiencies at scale.
90% of companies missed hiring goals in 2025. That's not a trend. It's a crisis. Your AI tools aren't the solution. They're part of the problem.
What Good AI Automation Actually Looks Like
There's a difference between a chatbot that reads resumes and a true computer use agent. A chatbot can keyword-match and spit out a ranked list. It can't schedule interviews. It can't update hiring systems. It can't send follow-up emails. It can't negotiate offers. Those are the real work. The work that actually fills jobs. A computer use agent is different. It operates like a human on a real desktop. It can click buttons, type in forms, navigate web applications, and manipulate systems the way you do. It's not just analyzing data. It's taking action. It can pull candidate data, update the ATS, send personalized outreach, schedule interviews across calendars, and track offers through to close. This is where AI automation in HR stops being a gimmick and starts being useful. It's not about replacing recruiters. It's about giving them an iron-clad assistant that never sleeps, never complains, and never makes a typo.
Why Coasty Is the Computer Use Agent You've Been Waiting For
Most AI tools in recruiting are wrappers around resume parsers. They're narrow, fragile, and limited. Coasty is built from the ground up as a general-purpose computer use agent. It scored 82% on the OSWorld benchmark, which is the leading independent evaluation for computer use agents. OpenAI's computer use scored 38%. That gap isn't a rounding error. It's the difference between an agent that can actually do work and one that mostly fails. Coasty controls real desktops and browsers. It can work in your existing ATS. It can integrate with Slack, email, calendars, and whatever systems you already use. It can run in the cloud or on your own infrastructure. You can even bring your own keys. There's a free tier. For teams that want to try before they commit, that's an easy decision. The point is that AI automation for HR doesn't have to be another expensive vendor. It can be a tool that genuinely works. It can be an agent that actually fills jobs instead of just telling you which ones to look at.
Stop buying AI recruiting tools that promise to fix bias but actually amplify it. Stop automating manual work with narrow bots that can't actually do anything. The future of AI in HR isn't resume parsers. It's computer use agents that can actually take action. Coasty is the only agent that's proven it can work at 82% on OSWorld. That's not hype. That's data. If you're serious about filling roles faster, cheaper, and without the legal risk, you need something that can actually use a computer. Not just talk about it. Go to coasty.ai and see what a computer use agent can do for your recruiting team.