Can AI Make Recruiting More Human? Kristen Habacht (Elly) on the Hiring Arms Race (39 min)
ai-bias-fairness ai-driven-innovation-economy ai-in-everyday-life ai-in-skill-development ai-in-workforce-disruption ai-literacy-public-awareness privacy-in-the-ai-era
- Release date: 2026-02-26
- Listen on Spotify: Open episode
- Episode description:
Recruiting is in an arms race: job seekers spray AI at every open role; recruiters crank filters to keep up. Nobody wins. ⚡ Kristen Habacht—CEO of Elly and former head of revenue at Trello (then Atlassian) and CRO at Typeform—thinks the fix isn’t more filters. It’s tech that actually learns, so recruiters can do the human work.Wade and Kristen talk about why most ATSs are “filing cabinets,” what “ICP for hiring” would look like, and why Elly never says yes or no to a candidate—only “did you see this? Is it important?” They cover the 1,000-applicants-in-24-hours reality, PLG in talent/HR, bias and AI screening, her take on AI “cheating” in interviews (“who really cares? It shows they know how to use the tool”), and why she’s giving away a lot of free usage instead of buying a billboard. Plus the story of the day she found out Trello was being acquired by Atlassian—and shoved her co-founder thinking he was joking.In this episode, you’ll hear:- 🎯 Why recruiting is broken for both sides and how the current arms race got here.- 📋 What’s wrong with today’s ATSs and why “ICP for hiring” could change the game.- 👥 How Elly keeps humans in the loop—and why the AI never decides anyone in or out.- 🔍 Why keyword search fails hiring (and what Trello’s early PLG hiring had to do with it).- 🤖 Whether AI in interviews is “cheating,” proctored interviews, and what might actually stick.- 🔗 Why TA and HR still don’t talk—and why that has to change.- 🌐 What a “universal job application” could look like—and where Elly is putting its $8M.Guest: Kristen Habacht — CEO, Elly.
Summary
- 🚀 Elly’s Origin Story: Kristen Habacht founded Elly to fix delightful-yet-broken recruiting, drawing from Trello/Typeform experiences where admins stifled hiring managers like herself.
- 📈 Spam Arms Race: AI-spammed applications (e.g., 1,000 in 24h) force harsh filters; Elly counters with ICP-building and human-like screening for unconventional fits.
- 🔄 Feedback Loop Gap: Hiring needs sales-like ICP learning from performance data; merging ATS/HRIS reveals true success traits, challenging biases.
- ⚖️ Bias & Human Loop: Audited AI highlights insights without deciding fits, reducing human biases and enabling deeper interviews.
- 🌐 Future Vision: Universal apps, AI fluency probes, and cheating tolerance signal PLG recruiting revolution ahead.
Insights
Can AI paradoxically make recruiting more human by surfacing non-obvious talent amid application spam?
Time: 10:20 – 13:09
Category: AI in Workforce Disruption, AI Bias & FairnessAnswer: Recruiting faces an arms race with AI-generated spam applications overwhelming systems, leading to rigid keyword filters that miss unconventional fits; Elly uses AI to build ideal candidate profiles (ICP) and enable deeper candidate stories, freeing recruiters for high-value human interactions. This shifts focus from rote admin to genuine connections. (Start at 10:20)
Why don’t hiring systems learn from performance data like sales ICPs do?
Time: 18:01 – 20:23
Category: AI-Driven Innovation Economy, AI in Workforce DisruptionAnswer: Traditional ATS track hires but ignore post-hire performance, promotions, and OKRs, preventing refinement of success patterns; integrating recruiting with HRIS data could predict fits better, challenge biases, and boost diversity by validating traits empirically. Companies like Zapier are starting this, revealing flawed assumptions. (Start at 18:01)
How can AI reduce bias in hiring without making final decisions?
Time: 20:45 – 23:01
Category: AI Bias & Fairness, Privacy in the AI EraAnswer: AI enables audits for bias compliance in states like CA/NY, outperforming unchecked human biases; Elly highlights resume/interview insights ranked by hiring criteria but leaves in/out decisions to humans, fostering thoughtful discussions and catching overlooked details. This keeps humans central while minimizing errors. (Start at 20:45)
What’s the smartest way to screen candidates for real AI fluency?
Time: 25:10 – 27:14
Category: AI Literacy & Public Awareness, AI in Skill DevelopmentAnswer: Generic ‘tell me about AI use’ questions yield rehearsed answers; probe specific projects to see if AI integrates into problem-solving, revealing true adopters versus superficial users like email editors. This mirrors hiring for nascent skills like PLG, emphasizing curiosity over rubrics. (Start at 25:10)
Will we soon stop caring if candidates ‘cheat’ with AI in interviews?
Time: 29:43 – 32:32
Category: AI in Everyday Life, AI in Workforce DisruptionAnswer: As AI becomes a job tool, proctored interviews may fade since using it demonstrates proficiency, akin to emailing vs. handwriting; focus shifts to identity verification and problem-solving walkthroughs, evolving from coding tests to natural conversation mocks. This accepts AI as augmentation. (Start at 29:43)
Could a universal job application, powered by AI, end the hiring noise?
Time: 34:24 – 37:30
Category: AI-Driven Innovation EconomyAnswer: Current siloed apps drown in spam; a standardized, AI-parsed profile like dating apps or Common App could match candidates/companies efficiently, especially for non-tech sectors absent from LinkedIn. Employers must accept shared criteria, with AI bridging disparate data without overhauling ATS. (Start at 34:24)