Since offering sponsorships on my website for the first time, I’ve talked to a lot more founders who are interested in sharing their product with the world via one of my blog posts. However, they’re a little surprised to hear that I’m not going to just take the money and smile while I show off a broken tool they call AI for recruiting. I want to see that it actually works before I ever share it with you. 

I am fascinated - nicest f word I have - by how bad some of this is. I think it has something to do with the fact that the people promising to fix hiring have never been in charge of hiring (or even adjacent). In the best cases, we have technologists who hired their own teams. Still not an expert. In the worst, we have people who have never worked even one corporate job. They heard some lecture about how hiring is broken in their MBA program and think it’s easy to fix hiring. 

Can they build something great? I guess it’s possible. But it’s not likely they are fixing anything in recruiting if they don’t actually take the time to understand the real problem here: our recruiting data is messy at best and nonexistent in most cases. 

Avoiding AI Failure: The Recruiting Data That Matters

The real reason why they fail is that what many of these founders see as administrative BS is actually the most important recruiting data. When I talk about recruiting data, I am not talking about time to fill or the other time sensitive metrics. Fact: Fast is not a reliable measure of success when you’re working with one of the most unpredictable variables in the world: people.

I’m talking about the data that defines what good means. Inputs most of these founders consider annoying administrative tasks like job posts and interviews are critical information that help us make decisions. That’s fixable - but it starts with you and no AI.  What does this person know how to do? In what context did they learn? If we start by teaching humans to clearly define what a good hiring output would be aka a job post, imagine all the possibilities for quality AI outputs. 

Think about it. How do you automate compensation if you don’t know the skills you need? How do you automate interviewing if you don’t know what you’re trying to learn about them? Better job posts unlock AI, but not the generic ones generated by the content blasters every ATS is trying to launch.  

Important note: To be clear, I am not talking about skills based job posts when I suggest defining a role. Skills based hiring is not an effective approach to categorizing roles or talent. It’s a bad investment. We’re marching into a world of unknowns when it comes to what people need to be good at. Creating an inventory now is a waste of time, IMHO.  

AI For Recruiting: Three Free (or Almost Free) Ways To Start 

Why? Well, most job posts are awful. That’s not going to change over night. However, that doesn’t mean you can’t use AI yet. There are easy, free ways to start with AI for recruiting that will make your life so much easier. 

No matter your company size, budget, team size, or anything else - here are a few things every recruiter should be using AI to do right now to save time so you can train your team to rewrite job posts and automate more work. 

1. Scheduling. If your team doesn’t have calendar links yet, what the hell are you doing? All that back and forth, be gone! I use Calendly. There are a million calendar options that allow for automated messaging and scheduling. Use them.
2. Notetaking. In the future of recruiting. conversations are data. Record and transcribe everything. Every interview, every chat, every thing. Be sure you’re using something with a license and not just uploading your stuff to ChatGPT. Every conversation you have should not be available to the whole internet. I use Fathom.
3. SEO Optimized Job Titles. It doesn’t matter how good your job posts are if no candidate ever finds it. Click here to try my free AI job title generator and learn how it works.  

What Recruiters Should Not Automate  

Here’s a list of things you should not be using AI for at this point. 

1. Broken processes. You can’t automate a broken process. 
2. Asynchronous video interviews. 70% of candidates don’t like them. I also don’t think they’re predictive of the ability to do much besides talk to yourself. Who in your organization needs to be good at that? 
3. Decision-making. In an AI world, you’re generating generic job posts with AI and candidates are using the same tools to tailor their content to perform well against your made up information. The risk? Both are mirrors of misinformation and then it’s almost a guarantee you will hire the wrong person. 
4. Write the first draft of your job posts. A good job post tells the truth and no machine can understand the truth and nuance of your posting without data. Bad data, bad output - and all the job post data these AI tools have is bad. 

Want to give your recruiters an introduction to AI? 

If you want to use AI to write a job post, you need to learn what good means first. My course is a how to and what not to do guide. You can buy it here.