The pressure is on recruiting teams to do more with less. In this environment, generative AI and its seemingly endless possibilities are an exciting development with unprecedented potential. But while AI can help with efficiency, it can also exacerbate issues such as bias. It’s important for hiring teams to have some AI tips in their pockets before using it.
Shockingly, one in three hiring efforts ends in failure (i.e., no-hire requisitions). A (big) reason for this is that recruiting teams, while responsible for final outcomes, don’t have intelligence on or control over all stages of the hiring process. And while job-post writing may seem like a natural fit for generative AI, it means giving up even more control.
Generative AI can do a lot of things well, but it can’t write inclusive, effective, on-brand job posts – not on its own, anyway. Job posts are marketing documents that can make or break your hiring effort. Using generative AI poses numerous risks, from non-compliance with local employment law to perpetuation of bias to inefficiency from lack of real-time job market data. And it still needs a human touch, which is why we created this list of AI tips. We hope you find them useful.
AI Tip #1: Fill in the blanks with branded messaging
This is obvious, but generative AI doesn’t know your company the way you do. It doesn’t know what your benefits packages are, the perks you offer, your stance on diversity, and numerous other mission-critical job post components. It also doesn’t understand your brand or how your teams actually work.
The first step when using AI is to fill the blanks in the job post it wrote for you (even if that post appears complete as-is). That includes things like a clear job location, a benefits package, and a statement on diversity, equity, and inclusion (DEI). Depending on the location of the job, pay transparency laws may also apply.
You could train AI on your own recent job posts with the latest brand tone, corporate-approved boilerplate content, benefits packages, et cetera. That would help, but only if the new job is similar to the older ones. Otherwise, you could end up using the wrong content for unique attributes like job type and job location.
It may feel like some of these things are assumed by job seekers, but they’re not. In fact, there’s a lot of confusion around job location these days, so it’s vital to state clearly where you expect someone to work (onsite, remote, or hybrid).
Also, job seekers are pressed for time too, so you should include all relevant information in the job post and avoid directing them to your careers page. And make sure you include a DEI statement in your job post. In a study of DEI statements, we found that job seekers perceive your company as more inclusive when you include one in the job post.
But a job post is more than a description of the job and list of benefits and perks. As the first and often only piece of messaging a job seeker sees from your company, it’s also an introduction to your brand. Think of it as a marketing asset wrapped in the clothing of a job post. Use it as an opportunity to spread the word about your brand, even with job seekers who don’t apply.
AI Tip #2: Make it unique
Saying that AI can’t write good job posts isn’t a knock against it because humans struggle with it too. In fact, there are parallels between how AI writes job posts and how humans do. They both look for examples on the internet and tweak them to their needs. (In other words, AI learns from what’s already out there and doesn’t imagine anything new.)
The problem with that approach is that it perpetuates the job post echo chamber where all job posts end up reading alike. (The more similar-reading job posts there are out there, the more similar-reading job posts will follow.)
The job post echo chamber is the reason why so many job posts have soft skill requirements like “excellent communication skills,” which is entirely subjective. Do we mean someone who can write emails, someone who can write blog articles, or someone who can speak to a room full of people? And are we really asking the job seeker to judge themselves? Soft skills don’t mean anything as part of a job post, but they can deter qualified job seekers. (Assess them in the interview process.)
The more unique content you include in a job post, the less your job post will sound like every other one out there. And that’s important because today’s job seekers don’t want to read vanilla job posts.
AI Tip #3: Personalize and humanize your messaging
Impersonal job posts don’t perform particularly well because today’s job seekers want to read something that speaks to them directly. Meanwhile, AI-generated job posts don’t perform particularly well because job seekers also want to feel like a human is speaking to them.
If you’re using generative AI to write job posts, you will likely need to personalize the messaging. By that, we mean speak directly to the reader by using “you” language and paint an accurate picture of the job so someone can see themselves in it.
You’ll also have to humanize the language, which is more complicated. Much like humans, generative AI learns from ingesting information. A generative AI tool’s knowledge comes from whatever dataset a human puts into it (e.g., the internet).
But a human’s knowledge comes from a much wider dataset – other humans, schooling, and a lifetime of experiences. Because an AI tool hasn’t lived a life as a human, it doesn’t know how to speak as a human. You’ll have to help it on that one.
AI Tip #4: Follow data-driven guidance
In the first of these AI tips, we point out that you’ll have to fill in a lot of blanks in any job post written by AI. In addition to company-specific content like a DEI statement, you’ll have to add (or edit) role-specific things like job title, requirements, and reporting structure. Unlike, say, benefits, these are things that AI would struggle to even guess at by crawling your website.
Pulling that thread a little more, it’s exactly those unique attributes that require real-time, data-driven guidance. If the title you choose is obscure, job seekers may not find it in searches on job boards. When the title seems more senior than the responsibilities, you may deter qualified candidates because of the confidence gap. And if the requirements are confusing (e.g., a wishlist), qualified job seekers may take a pass. The same thing can happen if you don’t include a reporting structure. It goes on and on.
And again, the quality of the job post dataset is crucial. If you trained AI (or a human, for that matter) solely on job posts proven to attract qualified, diverse applicant pools, it (or they) could do a better job. But generative AI tools train on all job posts, not just the good ones.
In the end, you’ll need to optimize your AI-generated job posts with data-driven guidance from a tool purpose-built for recruiting. Generative AI trained on broad-based datasets and making decisions based on probability isn’t enough. The reason Datapeople understands what makes an effective job post is because it combines training on millions of real-world jobs and hiring outcomes with data science and behavioral science.
Follow these four AI tips for inclusive, effective job posts
Remember, job posts have a measurable impact on hiring. Done well, they can attract a qualified, diverse applicant pool that translates directly into a successful hiring effort. Done poorly, they can result in few or no qualified applicants and a failed hiring process. And, potentially, damage to your employer branding efforts.)
When using generative AI to write job posts, include all the information job seekers want to see, in branded copy. Add unique content so that your posts don’t look like every other one out there. Personalize and humanize your messaging so you speak directly to job seekers as a human would. And optimize your post in a platform with real-time, market-calibrated guidance to make your post inclusive and effective.
If you’re looking for more information, we’ve got an article on using generative AI for job ads to check out. (We talk about actually testing AI-generated job posts in Datapeople.) Or you can watch Perspectives on Generative AI’s Transformation of Talent Acquisition, our webinar with leading industry analyst Tim Sackett. And if you’d like to see what optimizing a job post in Datapeople is like, you can also sign up for a demo.
