Wellness
Rewards
Recognitions
Learn more about Wellable's all-in-one
employee engagement platform
Do’s & Don’ts of Modern Hiring with Kat Kibben
Estimated reading time: 8 minutes
In this week’s episode, Geoff sits down with Kat Kibben, keynote speaker, bestselling author, and one of LinkedIn’s top voices on hiring. Drawing on 15 years working across HR technology, employer brand strategy, and recruiting education, Kat shares a no-nonsense playbook for key stages of the hiring funnel, from writing job postings that actually attract the right candidates to using AI screening tools without embedding bias.
Short on time? Here are the key takeaways:
- The real reason you are getting too many unqualified applicants is almost always the lack of specificity in job posting, not the volume of candidates in the market
- Candidates should apply if they meet 60% or more of a job posting’s requirements, because recruiters rarely expect 100% fit and most postings significantly overstate what is actually needed
- AI screening tools should be limited to clear yes-or-no questions at the initial stage and used to clarify ambiguous criteria in the maybe pile—machines should gather information, not make hiring decisions
- When companies use AI or bot-driven screening, candidates often are not told they are talking to a bot and do not know how to adjust, so these interviews reward the most confident performers rather than the best-fit candidates
- Bank of America’s AI upskilling framework centers on one principle—never trust the yes—reflecting a shift from teaching prompt skills to building the human judgment needed to evaluate AI outputs critically
Episode Summary
The Job Posting Problem: What Companies Are Getting Wrong
Most companies say they want to stand out in recruiting, but most job postings are extremely generic. Kat opens the conversation by explaining a search she ran across 500 job boards for the phrase “highly collaborative team player” and found it more than 15,000 times. That kind of generic language is not just unhelpful for candidates; it is actively breaking the AI matching systems recruiters rely on to surface qualified applicants.
The core mistake: Job postings describe what a candidate should be, not the scope and scale of what they will actually do. Kat’s argument is that the reason most companies reject candidates is not that they cannot do the job in theory but that they cannot do it at that company’s scale, in that industry, at that level of complexity. A posting that fails to communicate scope and scale will attract applicants who cannot make that judgment for themselves.
What to do instead:
- Replace generic competency language with specific scope and scale descriptors. Instead of “strong communicator,” say “manages cross-functional stakeholder communication across a 500-person organization.”
- Define the mandatory requirements before writing anything else. If you cannot explain what qualifies a candidate in plain language, your AI tools cannot either. Kat believes a large part of why 95% of AI projects fail is for this reason.
- Treat the job posting as the foundation for everything that follows: screening questions, interview structure, evaluation criteria. If the posting is vague, everything downstream will be too.
For candidates: Kat’s advice is to apply if you meet 60% or more of the listed requirements. Recruiters do not expect 100% fit, and candidates consistently screen themselves out for not meeting criteria that were never truly mandatory.
The LinkedIn Volume Problem: Why You Are Getting the Wrong Applicants
High application volume is one of the most common complaints Kat hears from recruiters. Her diagnosis is that most teams are blaming the wrong thing.
The real cause of unqualified applicant volume: A vague job posting tells AI matching bots that almost anyone qualifies, and surfaces job “matches” to a broad range of candidates who are not actually qualified. Recruiters receive hundreds of applications that technically meet the stated criteria but are a poor fit for the actual role. The posting is the root cause of the problem.
On fraud candidates: A growing share of inbound applications are not from real people. Kat acknowledges this as a compounding problem, one that specific scope and scale language helps filter because generic postings are far easier for bots to match against than specific ones.
The longer-term risk nobody is talking about yet: A bad job posting is an AI building block. Every vague posting feeds into matching algorithms that learn from it. Poor inputs now create poor AI outputs at scale later.
How to Use AI Screening Without Introducing Bias
Kat draws a clear line between where AI screening tools add value and where they create problems. Her framework is simple: machines should gather information to help humans make decisions. Machines should never make the decisions themselves.
Where AI screening works:
Stage 1 — Initial yes-or-no screening: Use bots only for questions with objectively correct answers. Kat’s standard is what she calls intern-level screening: a question that someone with zero recruiting experience could ask, receive a yes or no answer to, and make a yes or no decision on.
- ✅ “Do you have authorization to work in the United States?”
- ✅ “Are you available to work on-site in [city] five days per week?”
- ❌ “Do you have strong collaboration skills?” (Not a yes-or-no question. Not universally understood.)
Stage 2 — Clarifying the maybe pile: After narrowing 1,000 applicants to 200 candidates who could plausibly do the job, and 50 who could definitely do it, use bots to clarify the criteria you were unsure about in the 150 who do not fully meet requirements. This is where well-defined upfront criteria pay off: if you have been specific about what you need, generating targeted follow-up questions is straightforward.
Where AI screening goes wrong:
- Pre-screening bots are being deployed with questions that are subjective, industry-specific, or unclear, which means they are measuring how well a candidate interprets a vague question, not whether they can do the job
- Candidates are not being told they are talking to a bot, and most people do not know how to thrive in that interaction. The people who perform best in bot-screened interviews are often the most confident performers, not the best fit. As Kat puts it: whose job is it to sit and talk to no one? That is not what you want to measure.
- Speed-focused screening prioritizes moving candidates through the funnel quickly, but speed toward the wrong result is not efficiency
Kat’s recommendation for HR teams using video or bot screening: Tell candidates they are talking to a machine and give them guidance on how to approach it. Transparency improves the quality of responses and reduces the advantage of candidates who are simply good at performing confidence.
Navigating AI Change with Confidence: What Actually Works
The first wave of AI education focused on prompt writing. Kat’s view is that the field has moved past that, and the organizations doing it well have figured out something more important.
What Bank of America got right: Their AI upskilling program leads with one principle: never trust the yes. The skill they are building is not technical proficiency with AI tools but human judgment about AI outputs. In Kat’s framing, the quality of the prompt is not the problem. The quality of the human decision-making about whether the output is good is.
What this means for HR leaders:
- Upskilling programs should prioritize critical evaluation of AI outputs over tool-specific training
- Junior employees are especially at risk of over-relying on AI outputs without the contextual judgment to evaluate them. Pair them with experienced colleagues when experimenting with new tools rather than letting them learn in isolation
- Encourage teams to ask: do I actually like this? Is this actually good? That human layer of judgment is the skill that compounds over time
On resilience and navigating uncertainty: Kat ran her company from the back of a van for three years, not as a brand moment but as a practical constraint that stripped away the usual operating mechanisms. What she learned from that: there is no tool that solves all your problems. The skills that actually help are asking for help, changing direction without drama, and knowing when to quit. Treating quitting as failure has led to a cultural reluctance to change course that is particularly costly in an environment where conditions are shifting as fast as they are now.
Rapid Fire: Kat’s Straight Answers for HR Leaders
One thing HR leaders should stop doing immediately: Copying and pasting old job postings. If everything about the role has changed in the past six to twelve months, the posting has to change too. A recycled posting in a changed environment is not a job posting. It is a liability.
Most overrated hiring trend: Interview bots. Not because automated screening has no value, but because most implementations are not asking the right questions, which means they are not getting the right answers. Kat calls out the risk directly: you are materializing and creating consistency around bias, not eliminating it.
A team doing it right: Zapier’s recruiting team. Kat points to them as an example of organizations focused on results rather than speed, using follow-up screening thoughtfully and being transparent about their process.
Where to find more: Kat publishes a weekly blog at threeearsmedia.com, with a Tuesday post focused on one immediately actionable recruiting tip and a Friday letter on navigating work and life. Both are worth subscribing to.
Frequently Asked Questions
Using generic language that describes what a candidate should be rather than the specific scope and scale of the work they will actually do. Phrases like “highly collaborative team player” appear over 15,000 times across job boards and mean nothing to candidates or to AI matching systems. Kat’s recommendation is to replace competency language with specific descriptions of what the role involves at the level of complexity, industry, and organizational scale relevant to that position.
Yes, if they meet at least 60%. Kat’s experience is that recruiters rarely expect 100% fit, and candidates routinely screen themselves out based on job postings that overstate actual requirements. The decision to apply should be based on realistic assessment of fit, not on whether a checklist is fully satisfied.
Kat’s framework limits AI screening to two use cases: initial yes-or-no questions with objectively correct answers, and follow-up questions to clarify ambiguous criteria in the maybe pile. In both cases, the machine gathers information and the human makes the decision. Screening bots should never be used for subjective questions or to make final screening decisions, and candidates should always be told when they are interacting with a bot.
Two risks stand out. First, vague or subjective screening questions mean the bot is measuring something other than job fit, often the candidate’s comfort with ambiguity or their ability to perform confidence under uncertainty. The candidates who do best in these interactions are not necessarily the best fit. Second, without transparency about the screening process, candidates cannot prepare appropriately, which further skews results toward performers rather than the best qualified applicants.
Bank of America’s approach is Kat’s benchmark: center the training on judgment, not tools. The principle “never trust the yes” reflects a shift from teaching employees how to use AI to teaching them how to evaluate AI outputs critically. The skill that compounds over time is the human layer, the ability to decide whether an AI output is actually good.
Kat’s three principles for navigating uncertainty: ask for help, change direction effortlessly, and allow yourself and your team to quit what is not working without treating it as failure. In a fast-changing environment, the organizations that can course-correct quickly are better positioned than the ones that stay committed to a plan that is no longer serving them.
Full Episode Transcript
Geoff: Welcome to the Wellable Weekly Podcast, where we cover the key topics and trends at the intersection of well-being, technology, and HR. I’m Geoff, and today we have a very special guest. Kat Kibben is a keynote speaker, bestselling author, and top voice on LinkedIn on hiring. Kat has over 15 years working alongside top brands like Zoom, Monster, and Ollie. Kat teaches teams how to navigate change with confidence and build the kind of resilience that leads to better decisions. Their work has been featured in the New York Times, NPR, and Forbes, and they are the author of two books, including This Was All an Accident and The Bounce Back Factor. Kat, welcome to the show.
Kat: Thank you for having me.
Geoff: Your company is called Three Ears Media — there has to be a story there.
Kat: Three Ears Media is named after two dogs with four ears. I always clarify that because the first time I didn’t, I had a wild conversation with someone whose dog’s ear was amputated. I started my company in a quick moment, literally did a spin in my chair, looked over at my dogs under a blanket, saw three ears sticking out — a lab with one ear that stands straight up and a little Boston Terrier with two ears — and thought, what if I do Three Ears? The domain was available and here we are almost nine years later.
Geoff: Nine years. You’ve become such a prominent voice on the human side of hiring and recruiting. How did that happen, especially in the age of AI?
Kat: I have one of those resumes that looks confusing, but if you see where I ended up, it all adds up. I started on the technology side working for Monster.com, understanding the relationship between candidates, recruiters, and the technology they use. From there I worked with HR technology companies, then became managing editor of a recruiting blog for five years where all I did was talk to smart people and write about what they did. I believe that’s always been the most powerful thing in recruiting, because there is no education pipeline into recruiting. We can’t assume every recruiter knows the same thing about what works, what doesn’t, or the data that drives it. After that I did employer brand copywriting for some of the biggest companies in the world, then started Three Ears Media to create my own path of training and education. I believe making hiring more human starts with a better baseline, setting expectations so we all agree on what good means and know how to execute it.
Geoff: What are companies getting most wrong about job postings right now?
Kat: A lot of companies say they want to stand out and then do nothing different. It has never been easier to create something generic that says a lot without saying anything. A few days ago I searched a tool that allows me to search 500 job boards. I typed “highly collaborative team player” — the most generic cliche — and it appeared over 15,000 times. That creates a matching problem. If candidates are using AI to match against your posting and your posting is full of generic language, you get generic results. And candidates make one big faulty assumption: they decide whether they are qualified based on 100% of the job posting when recruiters don’t actually believe they need 100% of those things. I tell people: if you’re 60% or more qualified, go ahead and apply.
Geoff: What is the missing detail in most postings?
Kat: Scope and scale. The reason recruiters say no to people is usually not that they can’t do the job. It’s that they can’t do it at our scope and at our scale. Startup experience doesn’t always translate to enterprise. A highly niche industry doesn’t always translate to another. That scope and scale is how both sides of the equation can identify what they’re looking for and make the match. Most postings are lacking the specificity to help someone actually understand what is needed.
Geoff: You’re getting too many applicants. How much of that is the posting versus the platform?
Kat: I think people are blaming the wrong thing. If you’re getting too many unqualified candidates, there is a very high likelihood that you are not describing the requirements well. It is a matching system. Some percentage of your applicant pool is a bot that ran your job posting through a matching algorithm and submitted an application. That is causing a longer-term problem most recruiters have not identified yet. A job posting is an AI building block. If you cannot explain what qualifies a candidate in a way anyone could understand, how do you generate accurate screening questions? How do you define interview criteria? It starts with defining what you’re looking for, and we are lacking sophistication in that area. 95% of AI recruiting projects are defined as failures, and I think poor job postings are a significant contributing factor.
Geoff: What do you make of AI screening tools and how should HR teams be using them?
Kat: I have seen a lot of poor execution. I think automation can fit into screening in two places and be highly efficient. Efficient meaning they get accurate information that helps you make the decision, because machines should never be making the decisions.
The first place is the initial screen, but only on clear yes-or-no questions. Do you have a visa or authorization to work in the United States? Yes or no. That is universally understood. “Do you have collaboration skills?” is not a yes-or-no question and should never be in a bot screen. I call this intern-level screening: a question that someone with zero recruiting experience could ask, get a yes or no answer to, and make a yes or no decision on.
The second use is on your maybes. You narrow 1,000 applicants to 200 people who could do the job, but only 50 fully meet your requirements. The other 150 in that pile — use screening to go out and clarify the criteria you were unsure about. The beauty of this is that if you defined what you need up front, it is easy to generate follow-up questions and execute that at scale.
Geoff: What are the pitfalls of over-relying on screening technology?
Kat: Two dimensions. First, we use tools for pre-screening when they are not asking the right questions. It doesn’t matter whether it’s a tool or a person — if you don’t ask the right questions you don’t get the right answers. Most of these tools are programmed around a belief of what good looks like at the largest scale, but every use case is unique. Second, we are not telling candidates they are talking to a machine, and we are not telling them how to thrive in an interaction with a machine. We assume people know how to talk to a bot. They don’t. We struggle to talk to other people and we are people. If a candidate doesn’t know they are talking to a bot and doesn’t know how to navigate that interaction, the people who perform best are going to be the ones who are good at making up answers, good at projecting confidence. You are measuring someone’s ability to talk autonomously, not their ability to do the job.
Geoff: Workers are absorbing nonstop change. Layoffs, RTO, AI reshaping roles. What does navigating change with confidence actually look like?
Kat: Bank of America’s AI upskilling approach really inspired me. The number one principle is never trust the yes. I love that. The first wave of AI education was all about skills — prompts, prompts, prompts. But now I think we are realizing that the quality of the prompt is not the problem. The quality of the human deciding whether the output is good is the problem. And that is what you see on LinkedIn: bots writing comments that all sound exactly the same. The upskilling that matters now is building stronger behaviors that machines cannot replicate — judgment, critical thinking, the ability to decide if something is actually good.
The second thing is that it is all about people. I wrote The Bounce Back Factor before everyone was talking about AI replacing everyone. What I learned from running my company from the back of a van for three years is that there is no tool that solves all your problems. The fixes that work for people who live in normal situations don’t apply in unusual ones. What does work are tenants like: ask for help. Learn to change direction effortlessly. And sometimes, learn how to quit. Most people treat quitting like the big ugly Q word. But knowing when to change course is a skill, not a failure.
Geoff: Rapid fire: one thing HR leaders should stop doing immediately?
Kat: Copying and pasting their old job postings and thinking it will just work. You can’t tell me everything changed and then copy-paste the thing you used six to twelve months ago.
Geoff: Most overrated hiring trend?
Kat: The interview bot. The automated screening box. We think we are getting a better outcome when actually we are just materializing and creating consistency around bias.
Geoff: Any organizations doing it right?
Kat: Zapier’s recruiting team. Follow them, get on their newsletters. They are focused on results, not speed. Speed doesn’t matter if you are rushing toward the wrong result.
Geoff: Where can listeners find you?
Kat: I am the only Katrina Kibben in the world, so spell my name right and you will find me. My weekly blog is at threeearsmedia.com — Tuesdays I write about recruiting with one tip you can apply right now, no budget, no project required. Fridays I write a letter about life, because work-life balance is an illusion. It’s just you showing up over and over again.
Geoff: Thank you so much, Kat. And for everyone listening to Wellable Weekly, you can find us on Apple Podcasts, Spotify, or wherever you get your podcasts. Thank you.
-
Share
-
-
-
-
