Why Your ATS is Killing Your Recruiting Success (And What to Build Instead) with Steve Bartel, CEO GEM, Live on Shally's Alley
Shally's AlleyJuly 11, 202501:01:28

Why Your ATS is Killing Your Recruiting Success (And What to Build Instead) with Steve Bartel, CEO GEM, Live on Shally's Alley

Steve Bartel reveals why traditional recruiting tools are fundamentally broken and how AI-native platforms can revolutionize candidate relationships. From his engineering days at Facebook, Blizzard, and Dropbox to building one of the most innovative recruiting platforms, Steve shares the hard truths about why ATSs track requisitions instead of people and why bolt-on AI solutions miss the mark entirely.

Guest: Steve Bartel, co-founder and CEO of GEM

Website: GEM Platform

LinkedIn: Connect with Steve

Expertise: transforming recruiting through AI-native platform design and candidate relationship management


SUMMARY

In this episode we dive deep into the evolution from transactional hiring to relationship-driven recruiting, the ethics of AI decision-making in talent acquisition, and why every candidate deserves immediate, constructive feedback. Steve breaks down the massive blind spots most founders have about candidate experience and explains why treating recruiting like a product function isn't just smart business - it's the future of competitive hiring.


Key Takeaways


  • 70% of enterprise hires are already in their recruiting database - but companies can't unlock this talent goldmine because their systems suffer from institutional amnesia
  • Recruiters now manage 55% more requisitions than three years ago while facing 3x higher application volumes, creating an impossible workload without AI assistance
  • AI should elevate and rank candidates, never reject them - the moment AI makes hiring decisions, you cross an ethical line that undermines fair candidate treatment
  • Traditional ATSs track requisitions, not people - they were built for compliance, not relationship management, which is why recruiting CRMs require completely different data models
  • 30-50% of smaller company hires come from previous touchpoints in their database, proving that relationship nurturing beats constant new sourcing
  • The hard part of AI isn't the algorithm anymore - it's having sufficient data context to make intelligent recommendations, which requires native integration
  • Every applicant deserves timely feedback about their status - this should be table stakes, not a nice-to-have feature that most companies ignore
  • Recruiting requires evergreen personal contact information - unlike sales CRMs that focus on work emails, recruiting success depends on tracking people across career moves
  • AI can provide basic qualification matching at scale - but humans must handle nuanced decisions like citizenship requirements to avoid discrimination
  • Immediate feedback is crucial for candidate experience - three months later, constructive criticism becomes meaningless noise


Chapters

00:31 – Welcome & AI-Generated Theme Music Demo

03:23 – Steve's Background: From MIT Engineer to Recruiting Tech Founder

08:42 – The Dropbox Recruiting Experience That Changed Everything

12:17 – Why Traditional ATSs Are Fundamentally Broken

18:43 – The Birth of Recruiting CRMs and Salesforce Experiments

24:55 – AI Augmentation vs. Automation: Drawing the Line

31:57 – Ethics in AI Recruiting: Where Human Judgment Must Prevail

37:21 – Regulatory Challenges: GDPR vs. Trade Secrets

42:15 – Founder Blind Spots in Candidate Experience

48:13 – AI-Powered Qualification Matching at Scale

52:27 – Success Story: Enterprise AI Feedback Implementation


Sound Bites


"Any of us who got into recruiting, we're people people. We do it because we care about candidates and bringing in the right people. That's life-changing." - Steve Bartel


"If I had a time machine, the first thing I'd do is abort the ATS. It literally exists only to mitigate risk and tracks requisitions, not applicants." - Shally Steckerl


About Shally: ⁠srcn.co/me⁠

Powered by the WRKdefined Podcast Network.