A couple of years ago, this would have sounded far-fetched. Now it happens every day. A hiring manager asks an AI tool for a job ad and gets a fluent draft in seconds. Hundreds of applicants ask their own AI tools to tailor a CV to it. A screening tool ranks one against the other, and a shortlist appears.
Everyone saved time. Nobody checked the thing the whole process was built on: whether the job description described what it actually takes to do the job well.
In short: the job description is the seed of every hiring decision that follows, and if you don't check what you put out, you can't judge what comes back. Our free Job Description Analyzer checks yours against the skills and behaviors that predict success in the role, scores it out of 100, and gives you five fixes ready to paste.
When I launched the analyzer at our recent webinar on AI in hiring, he started with a job description that is a little vague and a little generic. On its own, that's not a disaster. But screening builds the shortlist from it, so the shortlist comes back slightly skewed: the right people a fraction less likely to surface, the wrong ones a fraction more likely to slip through. The interviews start from that weaker shortlist. The final decision is made on a foundation that drifted at every step.
No single stage looks like a failure. That's the problem. The small errors don't add up, they multiply, and when you're sifting thousands of applicants, a small drift moves a lot of people the wrong way. The good news is that it works in the other direction too. Tighten the start, and better inputs compound into better decisions.
Around half of HR teams now use AI to write job descriptions, and about three in four of them find it useful, according to Fosway research we shared at the webinar. It's quick and it reads well. But a general AI tool is a prediction machine. It writes the most plausible job description for a title, averaged from every ad it has seen. It can't know what success looks like in your team, in your market, this year, and because the result reads so well, you won't notice what it left out.
Every applicant, every screening score and every interview question is then judged against that text. If the brief was off, the quality you get back is off too, and nothing further down the process will tell you.
Think about the best person you've had in a role. What set them apart was rarely on their CV. It was how they worked: how they handled pressure, made decisions without all the facts, or won over a difficult stakeholder. Those are behaviors, and they're exactly what most job descriptions leave out, because tasks and qualifications are easier to write down.
They are. Figures shared at the webinar put AI use in applications at 39% of candidates, and rising fast. A CV or cover letter that once took real effort now takes seconds, so it tells you far less about the person who sent it.
Handing the pile to an AI doesn't fix that either. Martyn Redstone, an AI and HR compliance specialist, ran the same batch of CVs through three AI models every day for three weeks. On average a CV moved 2.5 places from one day to the next, and more than half of the CVs never made a single shortlist, on any day, from any model.
So a sharper job description isn't about writing better bait for AI-written CVs. Its job has changed. It's no longer mainly the thing candidates match themselves to. It's your definition of success: the standard every applicant should be measured against.
That's why we built a screening tool to sit right at the top of the funnel. It reads your job description, works out what good looks like for that role, and weights the score accordingly. Every applicant answers twelve questions, which takes about two minutes, and gets one fit score back inside your ATS. It's built on Wave®, the behavioral science behind our assessments, and it's far harder for an AI to complete on a candidate's behalf than a CV is to generate.
Put the two together and the loop closes. The analyzer makes sure your job description names what predicts success. The Screener measures exactly that in every applicant, so your first sift is based on the person rather than on whoever's AI wrote the best CV.
You paste in a job description, whether you wrote it or an AI drafted it. The analyzer matches the role to an occupation in the UK Standard Skills Classification, published by Skills England, and compares your text with the skills and tasks that occupation depends on. It also maps the role against Wave® to find the behaviors that matter most, and checks whether your job description asks for them.
You get a score out of 100, the skills and behaviors your job description is missing, and five fixes written for your role that you can paste straight back in. It reviews the job description and nothing else. It doesn't assess candidates or make hiring decisions.
My rule for AI in hiring is simple: let it do the heavy lifting, put a person where it's weak, and check what it gives you against something objective and validated rather than taking its word for it. The analyzer follows the same rule. The score is an AI-generated estimate, so read it as a signal of where to look. The fixes are ranked, so start at the top, keep the ones that ring true, and rewrite them so they sound like your organization. You're still the one deciding what the role really needs.
A job description is the first domino in hiring. Get it right, and every step after it gets stronger: the shortlist, the interview and the decision. That's why we made the analyzer free, with no catch.
Once your job description names what predicts success, candidate screening measures it in every applicant, and Wave Connect Hire carries the same definition of success from the shortlist through to the interview and the decision.
Try the Job Description Analyzer, or talk to us about screening every applicant against the same standard.