Full candidate pipeline? Here’s how to find signal with conversational AI

Jump Links placeholder, DO NOT REMOVE
- Trust, not volume, is the real constraint: 89% of UK and Irish employers believe they’ve turned away real candidates over suspected fake or AI-generated applications.
- Adding more human interviews hasn’t closed the gap: 95% of UK applications still never reach a conversation.
- Structured AI interviews help 63% of UK and Irish employers evaluate more candidates and 36% surface strong candidates they’d have missed.
- Employers and candidates agree on where AI belongs: only 5% of employers would let AI decide who advances, and candidates say knowing a human will review their responses is what makes them comfortable with the format.
Every CV in your pipeline might be lying to you, or it might just look that way. Either way, you’re left guessing.
That’s the bind recruiters across the UK and Ireland are stuck in right now. Almost nine in ten employers (89%) believe they’ve turned away a real candidate because an application seemed fake or AI-generated, according to Greenhouse’s 2026 Future of Hiring Employer Report.
That’s a signal problem, and it’s a bigger one than most hiring processes are built to handle.
As jobseekers lean harder on AI to polish and send out applications, employers are finding it harder to tell who’s actually right for the role, even with more applicants than ever landing in the inbox. Sourcing harder won’t fix it. What actually works is a process that can find the signal already sitting in front of you, and conversational AI screening is how a growing number of employers are starting to do it.
AI has made it harder to trust your candidate pipeline
Trust, not volume, is the actual constraint right now. Alongside that 89% headline figure, 31% of UK and Irish employers say they often or very often screen out legitimate candidates because an application looked fake or AI-generated. That’s roughly a third of employers routinely losing good people simply because they can’t tell a real application from a convincing one.
Candidates feel the squeeze too: employers report seeing applicants try new ways to get past screening, from overstating their skills to formatting CVs to try to game AI reviewers.
Why adding more interviews doesn’t fix it
The most common response has been to add another layer of scrutiny. Of the 90% of UK and Irish employers who’ve changed their hiring process because of fake or AI-generated applications, 49% added another human interview. That’s a reasonable instinct, but it doesn’t close the actual gap: The fall 2026 Hire Standard by Greenhouse found that 95% of UK applications never reach a conversation at all, and only 10.2% reach an assessment, so for the other 90%, the entire evaluation is a CV scan. UK recruiters sift through 183 applications for every hire.
Today, only three in 100 applications ever reach an interview, and adding more human interviews can’t close that gap.
– Daniel Chait, CEO and co-founder of Greenhouse
“People should stay accountable for every hiring decision, but AI can afford opportunities that people alone cannot,” mentioned Daniel Chait, CEO of Greenhouse.
How to build a candidate pipeline that surfaces real signal
Strengthening your recruiting pipeline means giving every candidate the same structured shot at a first conversation, and treating candidate pipeline management as a data problem instead of a guessing game.
Give every candidate a real first conversation
Most UK and Irish employers (90%) want applications to reveal more about candidates than a CV alone can show, and 48% would consider offering a live conversation or interview earlier in the process to figure out if a candidate is right for the role. The trouble is time: candidates wait an average of 18.6 days for a first interview, the longest single stage in the hiring process, according to Greenhouse’s Fall 2026 Hire Standard, while plenty of qualified applicants never get a conversation at all.
That’s the gap conversational AI interviews are starting to close, giving every candidate a real shot at a conversation before being screened out.
The real choice is a resume lost in a pile, or 15 minutes to tell their story. Every candidate should get the second.
– Ophir Samson, Head of Voice AI at Greenhouse
Scale consistency for better outcomes
Among UK and Irish employers already using AI interviews, 63% say they can evaluate more candidates than they could screen manually, and 36% have found strong candidates they would otherwise have missed entirely. That’s a meaningfully different outcome than simply adding another round of manual review: it delivers more reach and more consistency at the same time.
Keep humans in the loop, where it matters most
None of this is about taking recruiters out of the loop. UK and Irish employers are clear about where AI should stay in a supporting role: only 5% would let an AI agent decide which candidates advance.
Candidates feel the same way about accountability. Only 8% of UK and Irish jobseekers would refuse an AI interview outright, and 57% of those who’ve had one say it helped them show their skills, according to Greenhouse’s 2026 Future of Hiring Candidate Report. 51% say they want to know a human will review their responses before any decision gets made. Candidates aren’t rejecting the technology. They’re asking for a person to stay accountable for the outcome, which is what keeps most of them comfortable with the format in the first place.
That trade pays off for employers too: time saved through AI screening tends to get reinvested directly into candidates, most often through more time with shortlisted candidates and more interviews earlier in the process.
What this means for top of funnel recruiting
This kind of structured hiring pays off well beyond any single recruiter’s call. It gives top of funnel recruiting teams something they’ve never really had: visibility into the stage that determines everything downstream.
Right now, a third of UK openings close without anyone being hired: only 65.9% of roles close with a hire, and employers run about nine first interviews for every hire, according to Greenhouse’s Fall 2026 Hire Standard. A recruiter’s gut call earlier in the funnel isn’t data recruiters can point back to later. There’s nothing to analyse and nothing to show when someone asks why a strong candidate didn’t advance. When every screen produces a score tied to defined, role-specific criteria, TA ops finally has something real to work with: where candidates consistently struggle, where strong candidates are getting stuck and whether evaluation is holding steady across recruiters and roles.
Where to start: 4 ways recruiters can act on this now
A few concrete places to start, and a few new skills worth building along the way:
- Audit one high-volume role this week. Pick a req where application volume is already a known pain point and map out where candidates are actually getting lost today: the CV screen, the first conversation or the handoff to the hiring manager.
- Build a simple rubric before you pilot anything. Define 3 to 5 role-specific criteria you’d want every candidate scored against, then use that same rubric whether the first conversation is run by a recruiter or by AI. Consistency starts with the rubric, not the tool.
- Pilot structured conversational screening on one high-volume role. Tools like Greenhouse Voice AI let you test this without overhauling your whole process: run it against the rubric you just built and compare results side by side with your current screen.
- Set a human review checkpoint, not just a human interview. Instead of defaulting to “add another interview,” build in a specific point where a recruiter reviews AI-generated scores before anyone advances or gets declined. That’s what keeps the decision accountable without slowing the whole process back down.
Start finding the right talent
Recruiters were never short on candidates. What they’ve been short on is a reliable way to know, out of everyone already in the pipeline, who actually deserves a conversation first. Structured, AI-powered screening is how that gets solved, by building a clearer funnel where the right person and the right role can actually find each other, with a human still making the call.
For the full data behind this shift, download Greenhouse’s 2026 Future of Hiring Employer Report directly from this page.
FAQs
What is the main candidate pipeline problem today?
It’s rarely a shortage of applicants. The main candidate pipeline problem is a lack of reliable signal, no consistent way to tell who in an already full pipeline actually deserves attention first.
Why do so many UK and Irish employers think they’re turning away real candidates?
As AI-generated and fraudulent applications have become harder to spot, many employers have become more cautious, and that caution sometimes catches legitimate candidates in the process.
How does conversational AI screening create more consistent evaluation?
By scoring every candidate against the same role-specific criteria in the same structured format, rather than relying on each recruiter’s individual judgement call.
Does conversational AI replace human decision-making in hiring?
No. It’s designed to surface better signal on candidates. Humans still make the actual decision to advance or decline someone.
Will candidates accept an AI-led first conversation?
Most will. Only 8% would refuse one outright, and most who’ve experienced one say it helped them show their skills, especially when they know a human will review their responses.



