The Changing State of Talent Acquisition

The episode delivers solid, practical insights about human-centric AI versus human-in-the-loop approaches, and a three-pillar evaluation framework (impact, risk, complexity). However, the conversation lacks concrete data, metrics, or named case studies beyond the repeated Amazon example. The practical recommendations on building AI literacy (standing meetings, vendor QBRs, conferences) are sensible but relatively straightforward and somewhat obvious to experienced operators.

“human-centric AI solutions are actually designing the implementation of AI specifically to augment the human work that does need to remain human and automating the rest of the work that can be repeatable”

“by looking at these three different dimensions of AI use cases, I mean, yeah, you still need to ask them if they can support your 16 interview types. But by going this level deeper, you're actually going to get a little bit closer to what makes these capabilities, these solutions, work within your organization”

The human-centric versus human-in-the-loop distinction is useful and somewhat novel framing for the HR audience, and the three-pillar evaluation framework (impact, risk, complexity) adds structure. However, the core anxieties (AI bias, job displacement, risk aversion in HR) and solutions (build literacy, test tools, manage risk) are well-trodden in the broader AI discourse. The guest is thoughtful but not contrarian or genuinely first-principles in approach.

“instead, we could be using AI to move faster and smarter and keep our capacity open. It's like do I want to remain in the assembly line or do I want to step up and be more strategic?”

“we need to be evaluating AI differently than we evaluate different types of tech”

Kyle Lagunas is a research analyst at Aptitude Research with 15 years in HR/HR tech, giving him legitimate domain expertise and a bird's-eye view of vendor and market trends. However, he is positioned as a researcher and thought leader, not an operator who has built and scaled AI systems at a company or led AI implementation at enterprise scale. His insights are informed but observational rather than forged in execution.

“my job is to study innovation cycles in the world of HR and tech, and here at Aptitude we focus exclusively on trends that span human capital management”

“I've been in the biz for almost 15 years, so since I was three years old”

The transcript contains minimal concrete data or examples. The Amazon hiring algorithm bias is mentioned multiple times but dated vaguely ('12 years ago, nine years' - actually ~2018). The conversational AI use case for candidate concierge and employee inquiries is described generically with no named implementations, metrics, or results. The three-pillar framework is explained but illustrated only abstractly. Operator would struggle to find actionable numbers or proof points.

“the Amazon example of they built a machine learning algorithm that was going to help them screen candidate like or help them to funnel through applicants right, and we still hear about it only moving forward men, and that was like 12 years ago, nine years, it was a long time ago”

“our conversion rates for A and B candidates in this job type is like submission to acceptance ratio is 90%”

The hosts ask solid, thoughtful follow-up questions (e.g., on the Gartner hype cycle, the disconnect between executive support and adoption, the three-pillar framework). However, they rarely push back on claims or challenge assumptions. When the guest claims risk is often perceived rather than real, the hosts accept it without probing for evidence. The conversation is collegial and well-structured but lacks the productive tension and deeper interrogation of a truly sharp interview.

“I wonder if you know what some of you said. Kyle really resonates with me and perhaps with a lot of our audience, which is I've only been in this space for what? Five, six years now, graham and I feel like we've been talking about AI since I started”

“So what I'm hearing from you is almost a paradox, which is often what you find when you have powerful ideas, I think, which is this idea of human in the loop, which sounds incredibly human centric”