TaPod

The episode contains some substantive insights about fraud rates (3% baseline, 9% in redundant markets), the two-track labor market split, and the specific problem of late-stage reference checking. However, significant portions are devoted to product demonstration, casual banter about drinks and personal anecdotes, and repetitive explanation of the Xref.me platform that dilutes the insight-to-filler ratio.

“Over the last 3 million candidates that have been referenced by us, we have 150,000 legitimate fraud cases and 100,000 have been ignored.”

“In a standard market, 3% of candidates will fake their references. Today in administrative redundant markets where the pressure is high, the applications are high, fraud goes up to 9%.”

While the two-track labor market concept and the idea of early reference collection are somewhat novel for the recruitment podcast space, much of the commentary on AI-generated CVs and resumes, fraudulent candidates, and the problems with late-stage reference checking are increasingly common takes in 2024. The cinema/Imax analogy and the primary contribution - Xref.me - is a product feature rather than a genuinely contrarian insight.

“It's almost like the AI is canceling out the AI, right? It's happening in investment.”

“The sentence that crept into our sector probably 25 years ago and has been the bane of everybody's life. Sentence is references available on request. Right. So it means nothing is complete garbage.”

Lee Seymour is the founder and CEO of Xref, a company operating in reference checking and background verification for 14+ years with significant market penetration (11 million people referenced in ANZ). He has clear operational experience and scale, though as a founder speaking about his own product he carries inherent promotional bias rather than independent practitioner perspective. Relevant and credible within the talent acquisition domain.

“founder and CEO, uh, of xref”

“We started in 2010 and people said I was an idiot for trying to automate something that gets done manually.”

The episode provides concrete data points: 150,000 fraud cases out of 3 million candidates, 3% vs 9% fraud rates by market type, 100,000 ignored fraud flags, 6,000 Xref.me profiles built post-launch, and specific pricing ($27 AUD). Geographic specificity is mentioned (Australia, New Zealand, Europe). However, lacks named company examples of fraud incidents, specific client outcomes, or detailed metrics on the product's adoption impact.

“Over the last 3 million candidates that have been referenced by us, we have 150,000 legitimate fraud cases and 100,000 have been ignored.”

“In a standard market, 3% of candidates will fake their references. Today in administrative redundant markets where the pressure is high, the applications are high, fraud goes up to 9%.”

Host questions are surface-level and rarely push back on claims. Craig asks about fraud and AI concerns but doesn't probe contradictions or seek deeper evidence. Lauren's question about employer mindset shift is a softball setup for the guest's answer rather than genuine inquiry. The conversation devolves into casual banter about drinks and personal anecdotes rather than substantive follow-up. No productive disagreement or challenging of the guest's framing occurs.

“Let me throw a, let me throw a tricky one at you, Lee.”

“But some, um, employers and line managers, hiring managers, they'll be like, oh, I don't know about that. When do we, they're not going to put a bad one in there, are they?”