How to Screen Job Candidates Before You Ever Invite Them In
If you've ever hired someone who looked perfect on paper and then completely fell apart once they started the job, you're not alone — and it's not bad luck. It's almost always a gap in how the candidate was screened before they ever got a seat in the interview room.
Melissa Franks, host of the Opt-In podcast, has spent three decades interviewing people — from opening 21 retail stores in a single year with group interviews of 10-20 people at a time, to leading corporate restructuring teams evaluating hundreds of existing employees. Across that time, she's built a screening process that catches misalignment long before it becomes an expensive mistake. Here's how it works.
Start With a Real Application — Not an "Easy Apply" Button
The first mistake most small businesses make is skipping a proper application entirely. When candidates can hit "Easy Apply" on LinkedIn and auto-fill their information without reading the posting, you end up flooded with unqualified applicants — retail associates applying for styling sales roles, plant operators applying for an "operations manager" job that actually needed someone who knew Monday.com and Asana.
The fix: build an application that takes about 30 minutes to complete, with open-ended questions. That time investment forces a candidate to actually think about the role, the company, and whether they genuinely want it — not just react to a job title.
A well-built application does four things for you:
Flags typos and spelling errors — an early signal of attention to detail
Shows you how someone communicates in writing, which matters if the role requires it
Gives you a dataset to cross-check against their resume and LinkedIn profile later
Reveals early cultural fit by asking what stands out to them about your company's mission or values
Ask for a Video Submission — But Make It Optional
Requesting a short video response tests something a written application can't: verbal communication, professionalism, and whether someone follows through on a process even when it isn't required. Keep it optional on the submission form — whether someone chooses to do it at all tells you something.
Delivery matters less than you'd think. A well-prepared, scripted video from someone who isn't naturally polished on camera still signals they took the process seriously. A slick, off-the-cuff video from someone sitting in their car, dressed casually, often signals the opposite.
Watch for AI-Written Responses
One of the newer red flags in this process is content that was clearly generated by AI rather than written by the candidate. If someone copy-pastes an answer about their "top three career achievements" and it's still in third person, or reads oddly generic, that's a signal — not that they used a tool, but that they didn't actually do the thinking the question required. And if they'll copy-paste on an application, they'll likely do it on your work product too.
Use AI to Filter — But Always Check Its Work
Once applications come in, knockout questions (minimum years of experience, required location, specific credentials) can be run through AI to fast-track your shortlist. This works well — with one important caveat.
In one real example, a client's HR team built a rigorous AI-based candidate screening process using Claude. When the results came back for 20 applicants, they were wrong: two or three genuinely qualified candidates were screened out, while candidates the AI flagged as qualified weren't actually a fit. The issue wasn't the tool — it was that the AI had been trained on the recruiter's criteria, which didn't have the full scope and context that someone deeply familiar with the role would apply. AI is only as good as the expertise behind the prompt, and it can't replace judgment from someone who knows the role intimately.
Read the Resume — Actually Read It
Once you have a shortlist, the real diligence starts:
Typos and spelling errors signal low attention to detail on the one document meant to be a candidate's best first impression.
Achievements without data are red flags. Every real accomplishment has a number attached — revenue generated, costs saved, team size grown, projects delivered. If a candidate can't quantify what they did, they likely didn't do much of it.
Duration in role matters more than number of jobs. It typically takes about three years to build mastery at a given level. Someone who graduated in 2025 and is already claiming a director title hasn't had the lived experience the role requires — with the exception of genuine equivalency, like a director title at a much larger, more complex organization being roughly equal to a VP role at a smaller one. Use judgment, not just job titles.
Cross-Check the Resume Against LinkedIn
Put the resume and LinkedIn profile side by side. Do the dates match? Do the achievements match? Do the titles match? Small, natural gaps (like LinkedIn simply not being updated with recent roles) are normal. But contradictions — different achievements, different durations, different titles for the same role — are a yellow flag worth asking about directly.
Why This Matters
This process isn't about interviewing tens of thousands of people the way Melissa has. It's about consistency: applying the same rigorous filter every time, so that only genuinely qualified, aligned candidates make it into the interview room. Skipping steps here doesn't save time — it just moves the cost of a bad hire further down the road, where it's far more expensive to fix.
If you're building out a hiring process and want a second set of experienced eyes on your screening — or want someone to run this diligence for you — [book a free consultation call] to talk through what your next hire actually needs.
Book a free consultation with On Call COO →
Melissa Franks is the founder of On Call COO and host of the Optin Podcast, where she helps founders move from surviving to thriving in business.
