Is AI Empowering or Frustrating Your Hiring?

Here’s how a hiring process might well go: 

1. You draft a job description.  You chuck it in to ChatGPT (or similar) to get some more ideas.  That helped.  It’s better. 
   
   
2. You advertise the job using your favorite medium – job boards / recruiters / social media.  You ask for a resume and cover letter. 
   
   
3. Your candidates take the job advert, job description and their current resume.  They put it all into ChatGPT and ask the AI to make their resume look more like the advert and job description.  They might give Chat GPT a bit of leeway to ‘make stuff up’.  They also ask ChatGPT to draft their cover letter based on the job description, advert, and your firm values as espoused on your website. 
   
   
4. But not all candidates do this, and some do it more than others.  So you receive some ‘perfect’ applications and others not so perfect. 
   
   
5. You score these variable applications against the criteria you developed.  Maybe ChatGPT helped you develop the criteria based on the JD?  Maybe ChatGPT can even review the resumes and cover letters to help you.  Does it love the ones it actually wrote more than the non AI ones?   
   
   
6. You shortlist candidates.  You’ve done a bit better than a lottery.  You hope!  You phone screen the candidates.  At least they are human and can hold a conversation (oh and if you’ve asked them instead to create a video application you need to read how much this is a brand destroyer). 
   
   
7. You interview the candidates.  You ask the standard competency based questions based on the candidate resume.  Maybe you get ChatGPT to help you with this?  Unfortunately some of the candidates have also asked ChatGPT.  Does your AI generated question beat their AI generated answer? 
   
   
8. The candidates with the better ChatGPT answers get through to a final interview.  This one is less formal so ChatGPT doesn’t help the candidate much.  So if they can look you in the eye and hold a conversation about your favorite hobby they will satisfy you. 
   
   
9. You employ the top candidate based on interview scoring.  They had the best ChatGPT responses and you liked them. 
   
   
10. Chances they work out as you expect when they start work < 25%.   
    
    
11. That’s not including the 50% who decline your offer because they got a better offer elsewhere. 
    
    
    

What went wrong?  There was nothing objective in this process.  You give the job to those who use the best AI, are better practiced at interviews, and who look like you. 

Some simple fixes: 

1. Use a comprehensive skills test before first interview.  Assuming they meet a minimum standard (and if they don’t you’ve saved hundreds $$ not interviewing them) you don’t need to focus on this at interview. 
   
   
2. Use a quality personality profile before first interview.  Rather than standard competency questions you can be laser focused on the issues highlighted which might affect work performance, especially if you can compare their profile to an ideal profile for the role 
   
   
3. Since you’ve got great evidence already you can use half the interview to sell the benefits of them joining your team.  Remember who has the power at interview these days. Maybe less than 50% will take that other offer!

Giles Pearson  |  After 18 years as a partner with a large public accounting firm, Giles founded Accountests to help those recruiting accountants make better hiring decisions