AI is coming for talent acquisition
I sat in on a presentation recently where a very smart, very respected voice in HR technology made the case for fully automating talent acquisition. The argument was compelling, the data was real, and by the end, I was simultaneously impressed and unsettled.
The core thesis: AI has moved from assistant to agent to what he’s calling “super-agent” — technology that can listen to a business problem and deliver solutions without being explicitly told how. In recruiting, that means end-to-end automation. Screening, scheduling, interviewing, assessing, offering. The human as optional.
He used the self-driving car analogy. We started by automating the driver’s job. Then we realized the goal was never the driver — it was the passenger. So we built Waymo. The implication: maybe talent acquisition isn’t really “talent acquisition” anymore. Maybe it’s just a set of processes we haven’t automated yet.
I’ve been thinking about that framing ever since.
What’s right about it
The inefficiency argument is real. Companies now conduct 42% more interviews per hire than they did in 2021. The average time-to-hire has grown to approximately 42 days. Recruiting teams are managing more volume with fewer people, and a lot of what eats their time — scheduling, status updates, resume parsing — is genuinely automatable and probably should be.
The technology is also further along than most HR leaders realize. Agentic workflows that pull from multiple systems, identify internal candidates, build development plans, and send communications are not a 2030 projection. They exist now.
And the bias argument deserves airtime: human screeners carry bias into every decision. Structured, consistent AI screening at least applies the same criteria to every candidate. That’s not nothing.
What’s missing from it
The self-driving car analogy breaks down at a specific point: Waymo works because the passenger doesn’t need to trust the driver. They just need to arrive safely. The relationship is irrelevant to the outcome.
Hiring doesn’t work that way. Whether a candidate accepts an offer, shows up engaged on day one, and stays through the first year — these outcomes are shaped by how the process felt. Did someone treat them like a person? Did the interaction give them honest information about the role? Did they trust what they were being told?
An automated process can move faster. But it cannot do those things.
There’s a second gap: the metrics. Speed and quality in hiring are measured by employer outcomes — time to fill, offer acceptance rate, 90-day retention. 61% of candidates report being ghosted after an interview, up nine points from last year. Nobody is measuring what it felt like to be processed. Nobody is asking whether the people screened out deserved a closer look.
Automation built entirely around employer efficiency optimizes for employer outcomes. That’s not wrong. It’s just incomplete.
Where this leaves HR and TA leaders
The honest answer is that some of what we do in talent acquisition should be automated, and some of it should not. The work that’s repetitive, high-volume, and rule-based — schedule it, automate it, let the agent handle it. The work that requires judgment, relationships, and trust — that’s where the human must stay.
The risk right now is that the technology conversation is moving faster than the strategic conversation. Leaders are asking, “what can we automate?” before they’ve asked, “what should we protect?” Those are different questions, and the order matters.
AI should make recruiting faster, fairer, and less administratively brutal. It shouldn’t make candidates feel like they’re being processed by a system that has no interest in whether the match is actually right for them.
That’s the line worth holding.