What 200,000 AI interviews taught us about candidate experience
At HR Tech Las Vegas 2026, Workable CRO Sarfraz Ali revealed a surprising finding: when candidates were given the choice to speak with a human instead of an AI recruiting agent, nearly all continued with AI. Here’s what the data tells us about the future of hiring.
AI in hiring is a controversial topic, particularly among job seekers.
Candidates worry that automated systems are rejecting their resumes before a human ever sees them. Recruiters are concerned about losing the personal relationships that make hiring effective. And employers are trying to understand where AI actually improves the process versus where it simply adds another layer of technology.
But what happens when we look beyond the debate and examine how candidates actually interact with AI?
Workable’s Chief Revenue Officer Sarfraz Ali took the stage to share findings from the company’s AI recruiting agents, including insights from nearly 200,000 candidates who progressed to conversations with recruiters or hiring managers.
One statistic stood out: Just 0.1% of candidates requested to speak with a human instead of continuing their conversation with Workable’s AI Agent.
The finding surprised even the team that built the technology.
And it raises an interesting question: Have we been thinking about the candidate experience the wrong way?
What the data revealed
Workable launched its AI recruiting agents in early 2026 with a straightforward goal: help recruiters and hiring managers spend less time on administrative work and more time making informed hiring decisions.
But the scale of adoption quickly created an opportunity to examine something else: how candidates respond when AI becomes part of their hiring experience.
During his presentation, Ali shared the following results from Workable Agent.
Every application was evaluated as it arrived. Candidates identified for further engagement were typically contacted within minutes, rather than waiting days or weeks for an initial response.
And when those candidates were offered the opportunity to speak with a human instead of the AI Agent, approximately one in a thousand requested the handoff.
That doesn’t necessarily mean candidates prefer AI to humans.
As Ali explained during the presentation:
“This is not saying candidates prefer talking to a machine.
Instead, he suggested that the results reveal something about the shortcomings of the traditional application experience.
For many candidates, applying to a job means submitting a resume, waiting indefinitely, and potentially never hearing back.
Against that baseline, a timely, relevant conversation that acknowledges their experience may represent a meaningful improvement, regardless of whether the interaction is with a person or an AI agent.
Eight ways traditional recruiting breaks down
To understand the significance of these findings, Ali first examined the structural problems that have shaped recruiting for decades.
He identified eight common issues:
-
Batching: Applications accumulate before recruiters have time to review them, delaying responses and potentially losing strong candidates.
-
Recency bias: Candidates are compared with other applicants in the same batch rather than consistently against the job requirements.
-
Pipeline drift: Hiring standards shift depending on the quality of the candidates reviewed most recently.
-
Inconsistency: A candidate reviewed on Monday morning may receive a different evaluation than someone reviewed at the end of a long Friday.
-
Ordinary bias: Fatigue, resume formatting, personal associations, and other human factors can unintentionally influence evaluations.
-
Late discovery: Important details such as salary expectations, work arrangements, and visa requirements are sometimes discovered too late in the process.
-
Complexity: Organizations introduce additional screening stages and administrative processes to compensate for existing inefficiencies.
-
Limited passive sourcing: Proactively identifying and engaging qualified candidates is so labor-intensive that many teams reserve it for only their highest-priority roles.
Individually, these problems may seem manageable. Together, they create a hiring process in which the timing and order of applications can influence who receives attention.
A strong candidate might be overlooked because their application arrived at the wrong time, or because a recruiter simply didn’t have the capacity to investigate their qualifications further.
For candidates, that can mean missing an opportunity without ever having the chance to explain why they’re a good fit.
What changes when AI evaluates candidates as they arrive?
Workable’s approach begins with defining what the organization actually needs from a candidate.
Rather than comparing applicants against one another, Workable Agent uses an ideal candidate profile established through the intake process. That profile specifies the requirements, qualifications, and relevant experience for the role.
The Agent evaluates applications against those criteria as they arrive, helping recruiters identify qualified candidates without waiting for applications to accumulate.
When a resume doesn’t provide enough information, the Agent can engage the candidate directly, asking follow-up questions to clarify their experience.
Importantly, these conversations aren’t limited to a predetermined script.
During the keynote, Ali demonstrated examples of Workable Agent asking contextual questions, distinguishing between related but different skills, and identifying when a candidate’s experience was relevant but not an exact match.
Rather than relying exclusively on keywords or the confidence of a written response, the Agent can assess the evidence a candidate provides and document that assessment for the hiring team.
Recruiters can review the resulting scorecards, explanations, and conversation transcripts within Workable.
Human teams retain control over the criteria, workflow, and hiring decisions.
The goal is to give recruiters more complete information earlier in the process while giving candidates an opportunity to demonstrate qualifications that might not be obvious from their resumes.
Why the candidate experience was the biggest surprise
When Workable began developing its AI agents, the initial focus was on improving efficiency for recruiters and hiring managers.
The expectation was that candidates might receive somewhat faster responses, but that the overall experience would remain largely unchanged.
The results suggested otherwise.
Candidates were getting timely outreach, engaging in conversations relevant to their applications, and answering questions that might otherwise have gone unasked.
Instead of submitting a resume and hoping someone would eventually review it, qualified candidates could begin clarifying their experience within minutes.
For recruiters, that means receiving more detailed information before the first human conversation.
For candidates, it means having another opportunity to demonstrate their fit for a role.
This distinction is especially important in an environment where many job seekers are skeptical of AI-powered hiring tools.
The data doesn’t establish that candidates universally trust AI or that automated conversations can replace every human interaction. It does, however, provide evidence that candidates are willing to engage with AI when it offers a useful, responsive experience.
What this looks like in practice
The presentation also highlighted how organizations are applying these capabilities to real hiring challenges.
Belong, a UK nonprofit specializing in dementia care, employs approximately 1,500 people and operates with a recruiting team of just three.
Its hiring needs span care assistants, operational roles, and senior leadership positions, all within a competitive talent market.
Using Workable Agent, Belong was able to expand its reach without increasing recruiting headcount.
The organization filled a Head of Operations position in just seven days, from posting to offer, with the Agent configured in under 35 minutes.
The example illustrates how faster candidate engagement and more consistent screening can help smaller recruiting teams manage complex hiring requirements.
It also demonstrates the potential value of bringing qualified candidates into the process earlier rather than allowing them to sit unnoticed in an application queue.
The future of hiring still depends on people
The findings presented at HR Tech don’t suggest that human recruiters are becoming unnecessary.
They point toward a different division of responsibilities.
AI can evaluate large volumes of applications against defined criteria, identify gaps in information, engage candidates promptly, and support sourcing efforts at a scale that would be difficult for recruiting teams to manage manually.
Recruiters and hiring managers remain responsible for interpreting that information, building relationships, evaluating people more holistically, and making hiring decisions.
That was the central argument of Ali’s presentation: hiring works best when recruiters, hiring managers, and candidates are all positioned to succeed.
For years, recruiting technology has largely focused on making existing workflows more manageable.
The opportunity with AI is to reconsider whether those workflows need to operate the same way at all.
And perhaps the most significant finding from Workable’s first 200,000 AI-powered candidate interactions is that improving recruiter efficiency and improving the candidate experience don’t have to be competing priorities.
When candidates receive timely attention, relevant questions, and an opportunity to explain their experience, the benefits can extend across the entire hiring process.
Watch the full keynote to hear Sarfraz Ali walk through the data, demonstrate Workable Agent, and explain what these findings mean for the future of recruiting.
