Recruiting’s Grand Bargain: Rebuilding Trust in an AI-Soaked World
The idea for this piece wasn’t based on my first-person experience with the current state of hiring. Rather, it came from the moment I realized the "AI Revolution" in hiring had failed both its candidates and its recruiters.
While talking with a group of recruiters, I heard them speak at length about their desire for AI candidate detection. I tried to clarify what they meant: was it for detecting candidates using AI to misrepresent their experiences or to hide keywords in the whitespace in their resumes? No, they meant any candidate that used AI in any part of the application process. Recruiters were so frustrated and burned out by the effect AI tools have had on their experience, that all they wanted to do was create a bubble where they could avoid AI as if it never existed in the first place.
After dozens of these kinds of conversations, it's clear that the only revolution TA leaders have experienced is spinning in their chairs in frustration, or, at the very least, spinning their wheels while trying to accomplish the simplest of tasks. What’s worse, these challenges in the hiring system only seem to get worse by the week. I may be a technologist, but I am not here to smugly proclaim that some new algorithm has made recruiters irrelevant overnight. Just the opposite- I think both legacy platforms and hot new startups alike are doomed to fail if they do not break down the problems facing the recruiting and HR fields to first principles, challenge all assumptions we have about how we do business, and build solutions that reemphasize the human.
Some hard truths
Hiring as we have practiced it for the last two decades is irreparably broken. Recruiters are burned out; they’re being asked to do so much more while the tools they use only get worse by the week. All this while navigating stakeholders who are sure that there's a purple squirrel out there for them (if there are 2000 applicants, one MUST be perfect, right?). Meanwhile, candidates stall out or accept a role that didn't ask for their patience as the company tried to find an opening across six calendars for a seventh round panel interview. Nobody is happy. Everyone is mad. Inbox has crossed 150 and Slack is just sitting on 9+. Somehow those bottles of gin and pinot grigio just seem to get emptier faster and faster.
AI is not only failing to make good on its promises, it's actively making everything worse. Here's a hypothetical: you love donuts. Every day, you are lucky enough to pick the one donut that looks good out of a buffet of pastries presented every morning. But one day, the pastry chefs get smart to your impeccable taste and start to dress everything up as that perfect donut. Croissant? Looks like a donut now. Egg bites? Believe it or not, donut! Donuts? Still donuts, but you're so much more reticent to try one because you've lost all grounding about what is and is not a donut. You're confused, the fake donuts are ambivalent, and the real donuts are frustrated.
In a post-Mobley world, nobody knows what to trust. Cities, states, and countries all have their own regulation about using algorithms in hiring decisions. Running afoul has huge downsides for you and your company, so you're under immense pressure from legal to avoid algorithms at all costs. All that time you spent on mandated trainings for the tools your company paid six figures for? Wasted.
Why I named a blog after a search algorithm
In the nerd world I live in, semantic fit means moving beyond a simple search into a deeper, genuine understanding of capability, context, and potential. It goes beyond searching for distinct words or phrases to focus on a broader connection to concepts not in isolation, but in relation to everything around it.
As it relates to our jobs, semantic fit is the act of aligning what a person brings and what work actually requires. Not just ticking boxes, but a real congruence between a person's capability and potential and what a company needs. Rather than searching for or filtering by keywords typed on a resume, semantic fits are those that match what a person means, rather than just how they said it. It's a switch between "does this person list the five skills we asked for?" to "do we understand how this person's skills and experiences will enable them to thrive in our workplace long term?".
More broadly than these, it reflects my firm belief that we must adapt alongside the world, seeking alignment with that which helps us each flourish. How does our concept of labor fit into our lives, and how is that going to change as new technology is developed? I think a lot of what we are experiencing is grief for an analog world that no longer exists, applying physical systems of interaction to a digital world. Our mental model of hiring was straightforward; we had relatively little information about who we hired and decisions were made through human intuition and trust. In our digital world, there's so much noise and outside "influence" trying to sell us the idea of a perfect decision that we find ourselves stuck in analysis paralysis. The tools at our disposal have not helped, and have sold us the dream of quantifying human instinct and flattening the messy reality. Rather, we must face the world as it actually is, and build our systems to help uplift, augment, and embolden the imperfect humanity so critical to hiring decisions rather than smooth away its edges. Otherwise, we’re just reopening the req for the thousandth time.
Reframing our mental model
Hiring in the digital age has been framed as a filtering problem- how do you let the right people in the door so that you only have to talk to the most qualified for a role? How do we quickly reject the bad fits so we can focus on the good? This is a well intentioned approach, but as we discussed with those tricky donuts above, it's gotten harder and harder to tell what is a sweet treat and what is vegetable hiding in Boston Creme clothing. What we need is a reframe: I propose that we start to think of hiring as an information problem. How do we learn enough about candidates and the roles that we're hiring for so that we can match good fits quickly, and create a robust CRM for the future. When we think about it, job descriptions are just things we used to put in a newspaper so that people would walk in and hand us our resume, presumably followed up by a nice Q and A session. The problem is, this process hasn't been used since about 1998, so why are we still using the same artifacts to accomplish our goals?
I believe that it is critical we put down the microscopes we use to analyze candidates' skills and experiences. Hiring can seem like a farce because each side of this market is optimizing for things that data and experience tell us aren't actually the things that can predict long term success within an organization. I believe the only way forward is to de-escalate and strike a new grand bargain between the hirer and hired- how do we build better structures through which we can find and apply for jobs that fit our current skillsets and future aspirations, and how can we find the candidates best suited to thrive, not just those that can best craft a resume.
What this blog will explore
Through my posts, I'm going to try and decipher the present +1. Not quite the far flung future, but the state of hiring, the technology that helps or hinders us, and how we can improve our relation to both. I seek to help us all navigate the proximate challenges while keeping one eye fixed on the horizon. I will focus on what technology has the potential to help now, as well as how we should radically shift our thinking and take our favorite sledgehammer to the nonsensical processes we have derived for hiring and developing talent. I will try to mix in an economic perspective on technology as well as a bare minimum of philosophy as it relates to our work. How we spend our time and think about labor reflects so much about our sense of self, and it's important to ask questions about how we can better structure our systems to account for the breadth of humanity that exists. I will also touch on my work for PerfectHire and share a builder's perspective on what we're learning as I (quite literally) put my money where my mouth is and build something new.
In closing
I would love for this to be a dialogue, rather than feel like a crazy person given a digital soap box. If you think I'm full of it, tell me! If anything I'm saying is resonating with you, tell me AND I'll send you donuts. When building products, I try to interview as many users as possible in order to understand the problems they face, and I would like to take a similar approach here. Send me your worst applicant stories, the most cursed JD’s you’ve been sent, the most unhinged interview plans. At the very least, we can have a laugh about it together.
