Episode 181
Pavan Kumar on How Eightfold Uses Agentic AI to Hire Smarter
Eightfold’s Pavan Kumar shares how agentic AI moves beyond task automation to conduct structured interviews, screen candidates at scale, and free recruiters for high-value work. Six hundred interviews in six weeks reveal where AI excels—and where human judgment still owns the decision.
Episode Key Takeaways
Agentic AI is not automation; it’s a thoughtfully designed workflow where the system handles operational tasks while humans retain decision authority. The AI Recruiter conducts tailored interviews based on each candidate’s resume and the role, asks smart follow-up questions, and surfaces structured summaries—but the recruiter still makes the thumbs-up or thumbs-down call.
Pavan’s team has conducted over 600 interviews across 35–40 roles in just six weeks, with both recruiters and candidates reporting exceptional feedback. Candidates appreciate the ability to interview at their convenience (weekends, evenings) and even request second chances, while recruiters gain consistent, rubric-based data they can absorb in three to four minutes instead of reviewing hour-long conversations.
Scheduling remains harder than interviewing. Multi-stakeholder calendar coordination defeats most automation logic; humans can negotiate meeting moves in ways AI cannot yet replicate. Conversely, sourcing top talent still demands human creativity, relationship-building, and persuasion—areas where recruiter ingenuity outpaces algorithmic recommendation.
Candidate rediscovery and proactive talent engagement represent the next frontier. AI should surface hidden talent already in the organization—former employees, offer declines, internal silver medalists—and recommend outreach campaigns or community events before the recruiter even asks.
Trust is the foundation for adoption. Start with a small group of optimists, measure outcomes rigorously, collect feedback from early users, and ensure candidates can opt out of AI entirely if they choose. Explainability, human-in-the-loop design, and external audits (including EEOC and OFCCP guidance) are non-negotiable.
Frequently
Asked
Questions
What is agentic AI in recruiting?
Agentic AI is a workflow where the system performs operational tasks—like conducting interviews, summarizing feedback, or opening requisitions—while humans retain decision authority. Unlike generative AI (which produces text), agentic AI takes actions. The AI Recruiter, for example, conducts tailored interviews, asks follow-ups, and surfaces structured summaries for recruiter review; the recruiter decides who advances.
How many candidates has Eightfold's team screened with AI Recruiter?
Since launching roughly 1.5 months before this recording, the team conducted 600+ interviews across 35–40 roles spanning the US, India, and EMEA. Feedback from both recruiters and candidates has been phenomenal, with candidates praising the human-like experience and flexibility to interview at their convenience.
Where does agentic AI struggle in recruiting?
Scheduling interviews across multiple stakeholders remains difficult; AI cannot negotiate calendar moves the way humans can. Sourcing top talent also resists automation—especially in competitive markets where relationship-building, persuasion, and creative outreach matter more than algorithmic recommendation. Offer negotiation in complex, high-touch scenarios also requires human judgment.
How do you ensure responsible AI in recruiting?
Build human-in-the-loop checkpoints so AI never makes decisions—only recommendations. Ensure explainability so you can audit why a candidate was recommended. Conduct internal and external audits, maintain an AI ethics council with regulators (EEOC, OFCCP), and allow candidates to opt out of AI entirely. Measure candidate experience continuously and address grievances transparently.
How should TA leaders pilot agentic AI?
Center everything on trust. Validate the system is well-designed and reliable. Start with a small group of optimists on both the business and recruiting side to avoid early disappointment derailing enterprise adoption. Measure outcomes rigorously, build custom dashboards if needed, gather feedback from users and candidates, then scale rapidly once confidence builds.