AI has changed hiring for good, is it for the better or for the worse? In this article Will looks at some of the biggest problems in recruitment at the moment, with a particular focus on AI.

Over the years, I’ve held leadership roles including Head of Data Science & Engineering, Tech Lead, and CTO, leading teams and driving innovation. I’ve hired for critical technical roles—Data Scientists, Software Engineers, Data Strategists, and others—and built high-performing teams across organizations. I’ve seen both sides of the coin, as a hiring manager and as a candidate, and I won’t be the first to say this, but it sucks!

Well, this absolutely rubbish process is not a people problem. One of my favorite parts of hiring is meeting so many amazing people and getting to know their motivations, life stories, and work experiences. I’ve never not learned something new during a hiring process. Recently, however, the good side of hiring has been pushed firmly—and quite abruptly—into the back seat. Why? Let me tell you.

AI Is Eating the World! 🌍

It’s here to stay, and it’s now part of hiring on both the candidate and recruiter sides of the process. As a founder of an AI-based product, I’m super excited about the opportunities AI is bringing to the world, but as a hiring manager, it’s made me pull my hair out on a regular basis! 😅

In the last year when hiring for software engineers, I’ve seen first-hand how profoundly and irreversibly AI has reshaped the hiring process. Turns out creating memes, writing code, and creating somewhat relevant blog header images weren’t the only things it was good for! Who would have figured?

Here’s a snapshot of some of the problems AI has created for me:

• 📝 Every candidate has a perfect cover letter that matches the job description precisely, with their work experience conveniently aligned with every requirement.
• 🖋️ CVs are often written or edited by AI to fit job descriptions.
• 📈 The number of applications per job has exploded. Applicants can now use AI SaaS apps to apply for hundreds of jobs a day automatically.
• 📋 Take-home tasks have now mostly become just a signal of commitment to the role.
• 🤔 AI usage during remote interviews to answer questions.
• 🤝 Assessing how well a candidate collaborates with AI in practice.

What This Means 💼

Cover Letters Are Dead ☠️

They haven’t been the most reliable signal in a long time, but now, a cover letter perfectly tailored to the job description is commonplace. Either way, it’s just a signal of who decided to use ChatGPT to write theirs. Some candidates polish the generated cover letter a little, but honestly, who cares? At this point, it’s useless. Cover letters are a dead signal, and we should abandon them as a principle.

Keyword matching is dead 📉

Many candidates are using AI to adjust their CVs to match job specifications. The result? Keyword matching is dead. Every single keyword is being added to CVs and now context is all that matters. And thank goodness—it was a completely ridiculous way of screening. The quicker it dies, the better.

The Total Time to Scan Through CVs and Applications Has Increased ⏳

The number of applicants per job has skyrocketed. This would be great if it meant more qualified candidates, but anecdotal evidence suggests the number of truly qualified candidates hasn’t changed - just more irrelevant or poor applications. The haystack has just gotten bigger. For instance, a recruiter I spoke to mentioned a Junior Software Engineer role in London that had to be closed within 24 hours after receiving over 1,000 applications. The worst part? “There’s no way I have time to review all these applications,” they said. “I’ll just go through enough to create a shortlist and reject the rest.” This means you’re now actually spending far less time with the good candidates, and most of your time sorting through the chaff! 🤯

Take-Home Problems are Now Mostly Just Indicators of Commitment 🏠

The days of sending a take-home problem, marking the answers, and creating a shortlist based on scores are over. With tools like OpenAI’s o1 models and Cursor’s Agentic Composer, it’s unlikely candidates will provide completely incorrect answers. Take-home tests now primarily show that a candidate is motivated to get the job and can either do the test themselves or use AI effectively.

If you use a take-home test, its value lies in asking candidates to explain it in detail later, especially in person, where interference is minimal.

Remote Interviews can't always be trusted 📹

It might sound far-fetched, but I’ve personally witnessed candidates using AI to answer questions during remote interviews. Some tell-tale signs include repeatedly asking to repeat the question, significant delays before responding, looking off to the side, and the unmistakable tone of ChatGPT-like answers. This technology is only getting more advanced, so in-person meetings will gain even more importance in the future.

AI Skills need to be evaluated too 🤖

You might think a candidate who avoids AI tools during the process is promising. However, while it’s nice to see someone’s unaided capabilities, what’s even more impressive is a candidate who knows how to use tools like Cursor or GitHub Copilot effectively. In 2025, being a 10x engineer means mastering these time savers.

Also, avoid questions that focus on things candidates would realistically use AI to handle. Instead, focus on questions that assess their ability to evaluate AI’s outputs or tackle problems where AI is less capable. 🛠️

So, What Are We Going to Do About It? Enter Squirrel 🐿️

At Squirrel, we want to take the pain out of hiring, reduce reliance on internal recruitment teams, and put the power back in the hands of hiring managers. 💡

Fun fact! Did you know that “Purple Squirrel” is a term for a perfect candidate for a job. Our dream at Squirrel (which also happens to be purple) is to find as many as possible!

We’re Building an AI Agent for Hiring! 🤖

• ✍️ Build job specifications, scorecards, and interview questions in minutes. No back and forward with internal HR Teams
• 📊 Deal with the influx of applications by screening CVs contextually against job specifications with customizable scorecards. No more keyword matching!
• 🎤 Push candidates through a live voice-based interview with our AI Agent to filter out unsuitable candidates early. Fewer very bad candidates making to your desk
• 📨 Provide detailed feedback to rejected candidates, even at the CV screening stage. You'll no longer be that guy…
• 🗓️ Manage the hiring process end-to-end, including scheduling interviews and tracking candidate progress. No more missed interviews or mis-communications

This Is Just the Start 🚀

Our dream is that hiring will be completely hands-off—apart from the most important part: spending high-quality time with candidates! With Squirrel, we aim to automate the tedious, time-consuming aspects of hiring so managers can focus entirely on meaningful interactions that build trust and uncover potential. By streamlining the process, Squirrel ensures hiring is efficient, candidate-friendly, and centered on what truly matters—connecting with great talent.