what 4,000 CVs taught our TA lead

Daniella Angel has spent nearly fifteen years in talent acquisition, agency and in-house, hiring everyone from entry level to executive. For most of those years she’d read every CV herself. 

Daniella had tried AI scoring on and off (with other tools) but had written it off. Then one role brought in 3,994 applications.

Now, Daniella runs AI candidate screening through Employment Hero’s AI Recruitment Agent powered by Hero AI. And her time on CV review and screening calls has gone from 13 hours a week down to two.

We sat down with Daniella to talk about how she got there, what she’d tell someone starting out, and the places she still won’t let AI near.

What is AI candidate screening?

AI candidate screening uses artificial intelligence to evaluate and rank job applications against criteria the employer sets, so a busy recruiter can review a shortlist rather than the full pile. 

In practice it usually covers three things: scoring and ranking every applicant as they apply, running a structured first-round interview, and summarising the results for a human to review.

The important part is what it doesn’t do. AI screening does an amazing job of narrowing the field. But a person still decides who gets hired. And that’s not just best practice. Since February 2026, Articles 22A to 22C of the UK GDPR have set out what you owe a candidate when a significant decision is made by automated processing alone: information about the decision, a chance to make representations, the ability to obtain human intervention, and the right to contest it.

See the ICO’s guidance on automated decision-making and profiling if you’re putting a policy together.

“Start with the role that broke the habit…”

“We were hiring a talent acquisition specialist for our Canadian team,” explains Daniella.

“It ran from January to March, six to ten weeks from start to finish, and it brought in nearly 4,000 applications. 3,994, to be exact.”

How much of your time did that take, and when were you doing it?

Daniella explained that she ended up spending approximately 60 hours getting through that CV pile. From CV reviews and screening calls to all the prep and admin around them, it was herculean.  

“It was one of six roles I was running at the time, while leading a team across three time zones. So often I ended up reviewing CVs in the evenings, because my days were full of calls. Nobody expected me to, but I didn’t want to leave applicants waiting.”

What was the worst part?

Not the hours, as it turns out. What stayed with Daniella was the awareness of everyone waiting, both applicants, and the team with the role to fill.

“The thing people don’t see is the queue. When you’re working through a pile systematically, the people at the bottom might be sitting there for days or weeks waiting on you. That’s a constant anxiety ticking over in the background.”

And there’s a structural problem underneath that. You can’t confidently shortlist until you’ve seen everything, which means the pile sifting isn’t optional.

“It’s a bit like shopping. You want to scan the whole shop before you decide what to buy, so you know your shortlist genuinely is the best of the pile. With 4,000 applications, that luxury disappears.”

You’d used AI screening before and didn’t rate it. What changed?

This is the part most HR professionals will relate too hardest: Daniella didn’t arrive at AI candidate screening as an enthusiast. She arrived as someone who had tried it and written it off.

“The tools I’d encountered before were early in their evolution and the scoring was usually inaccurate. Previously, I felt confident I could ignore what the AI was suggesting, because my own judgment was better.”

Eventually she stopped looking at the scores altogether. 

But then two things changed. The technology moved on a long way, and more importantly, so did what she expected from it.

“I don’t expect it to decide who gets seen. I expect it to point me in the right direction early. It’s the difference between being shown the aces up front and flicking through the whole deck.”

What AI candidate screening actually changed

Walk us through what happens now.

Applications are scored against the job description as they arrive. Daniella has configured hers so strong matches are shortlisted automatically and invited to record a short interview with the AI Recruitment Agent in their own time. She reviews those profiles and watches those interviews.

The headline difference is time: CV review and screening calls used to take her up to 13 hours a week, they now take just two.

“That’s gone into the parts of the job that actually need me. Talking to people. Briefing hiring managers properly. Thinking about whether someone will thrive in the team rather than just clear the bar.”

For a role like the Canadian one, she reckons it’s the difference between a six-to-ten-week slog and something that could realistically wrap in one to two.

That gap has a price attached, too. The median UK time to hire is 40 days and the median cost per hire is £6,125, and neither figure improves while the pile sits unread and unsifted. A role that drifts for an extra month costs you twice over, once in recruitment spend and once in the work nobody is doing.

What surprised you most?

Daniella expected automation to put distance between her and candidates. It did the opposite. Automating the front of the process moved her first human contact earlier rather than later.

“The AI interview replaces the faceless CV as the first real contact. Candidates get their faces and their answers in front of me sooner, in their own time.”

It’s also changed who you end up meeting.

This is the point Daniella wants to say louder, for the people at the back

When she was reviewing everything herself, her week contained a fixed number of screening slots. That scarcity forced hard cuts, and the cuts weren’t always about capability.

“People got cut easily, and plenty of them got cut for being poor CV writers rather than poor candidates. Now I’ll look at someone I’d have passed over before, because I can hear them rather than read them.”

Now, a candidate who writes a weak CV but interviews well finally gets a look.

Has the scheduling burden changed?

Enormously, and Daniella says it’s the most underrated part.

“There’s no diary wrangling. Candidates record whenever suits them, which matters a lot when you’re hiring across three time zones. I’m not losing a week to finding a mutually convenient half hour.”

What’s the candidate reaction been?

Daniella is careful to say it’s mixed. While she’s experienced the positive impact of AI screening firsthand, she also understands why the people on the other side of the fence might still be wary as it feels so new. 

“It doesn’t land everywhere. In some countries where we’ve been hiring software engineers, they’ve been lukewarm on it… But sales and marketing candidates are the complete opposite. They are curious and ultimately they really rate the experience.”

Which makes a certain sense. Long before any of this existed, sales candidates were recording introduction videos and sending them to recruiters unprompted. 

Some people would far rather show you who they are than hope you read their CV properly.

Employment Hero customers are seeing great results from AI screening too 

• Thermosash Group reports 99% faster screening and 90% less recruitment admin. 
• All About Caring saves around 15 minutes per candidate who applies, largely because the agent drafts the screening questions and the team can watch the recordings back.

Results are based on case studies and may vary depending on hiring volume and role complexity.

How the AI Recruitment Agent works 

Daniella’s process is one way of using the AI Recruitment Agent. But here’s the full picture, so you can work out which parts matter to you. 

Scoring and ranking. Applicants from every source are evaluated and ranked to show how well they match your hiring criteria, so a strong candidate from one job board doesn’t get buried under a weaker one who applied elsewhere.

Job description and rubric generation. Describe the role or upload a draft, and it produces the job description, recommends a salary range and builds the scoring rubric. You edit all of it. This is the step Daniella says is worth slowing down for.

Screening calls. Candidates who look like a fit but are missing information get an AI screening call to fill the gaps, rather than being discarded for an incomplete application.

Auto-shortlisting. Strong matches can progress to your shortlist automatically. You build the hiring stages, and you choose exactly when and how the agent is involved, whether that’s a first-pass screen or a deep-dive interview.

AI video interviews. The AI interviewer drafts structured questions from your job criteria, and each candidate gets an invite to complete their video interview in their own time. It probes for more detail and asks smart follow-up questions, so everyone gets a consistent experience. You can edit, delete or add your own questions and adjust the scoring criteria before the invite goes out.

Interview insights. Summaries, key strengths and potential areas to probe for every interview, with sentiment analysis and performance notes.

One-click onboarding. New hire data moves straight into your HR platform without re-typing.

And there are two ways to get it

Included in your HR plan. See plans for more details. 

Standalone, just hiring, on its own. You don’t need to move your HR system or change how you hire with the Standalone Recruitment Agent: find out more.

Seen enough? Book a demo and we’ll walk through the options on a role you’re actually hiring for.

Daniella’s advice on getting started: AI screening 101 

If someone’s reading this and thinking about trying it. What should they do first?

“Sort your criteria out before you switch anything on. The scoring is only as good as what you point it at, so if your job description is vague, your shortlist will be vague. Spend the extra time getting the role set up properly when you post it. That’s where the real work is.”

What would you insist on from any AI screening vendor?

“Customisation, first and foremost. At Employment Hero, we write our own interview questions and set our own scoring criteria. The job description and scorecard candidates get matched against are mine. I created them and I still own those guardrails.”

What about candidates who score badly? 

“Don’t treat the score as a decision. They don’t get auto-shortlisted, but the process isn’t blind. I’ll still look at the CV, I can just do it faster because I can see quickly that it isn’t a match. If someone’s interview invitation expires or something looks off, I’ll check in with them directly and can still offer a conversation.”

Where doesn’t AI screening belong?

“Executive search or headhunting a senior leader. That’s relationship-led and it should stay that way. And as I said, not every market responds the same way. You have to be intentional about where you place it rather than switching it on across the board.”

If a candidate asked why they’d been rejected, could you tell them?

“Yes, and I always will where I can. Between my own judgment and the system’s scoring, I can give clear, objective feedback oncwhere someone didn’t meet what we were looking for, and what we did like.”

What’s the objection you hear most from other talent leaders?

“People say: ‘Won’t this just replace human judgment with a black box score?’ And I understand it completely, because it’s natural to assume it isn’t customisable and it’s all happening behind the scenes. But that isn’t the case. I set the criteria, I set the job description…the AI is determining how closely someone matches what I’ve asked for. It doesn’t replace human judgment, it gives the human a fairer, more consistent starting point.”

And to someone who thinks AI screening is fundamentally unfair?

“I’d vehemently disagree, and I’d say the opposite is true. Human screening is inherently flawed. We all bring subconscious bias without realising it.. As well as fatigue and affinity bias to every stack of CVs. And ultimately, you can’t audit someone’s brain.”

“AI screening isn’t perfect. But when it’s properly designed it enforces evaluation based strictly on skills and competencies. The real advantage isn’t that an algorithm is immune to bias. It’s that, unlike human intuition, an algorithm can be audited, measured and continuously corrected.”

AI candidate screening: Common questions  

How does AI candidate screening work?  

Applications are scored and ranked against criteria you set as they arrive. Strong matches can be shortlisted automatically and invited to complete a structured first-round interview in their own time. You review the shortlist, watch the interviews and make the hiring decision. 

What happens in an AI screening interview?

The candidate receives an invite and records their answers to structured questions whenever suits them. The AI interviewer probes for more detail and asks follow-up questions based on what they say, rather than reading from a fixed script. It replaces the telephone screen rather than the interview itself, and a person watches the recording.

Does AI candidate screening reject candidates automatically?

It shouldn’t. With Employment Hero’s AI Recruitment Agent you build the hiring stages and choose exactly when and how the agent is involved. 

It’s also worth knowing where the law sits. Under Article 22C of the UK GDPR, if a significant decision is made by automated processing with no meaningful human involvement, the candidate must be able to obtain human intervention and contest the outcome. Keeping a person in the loop is the simplest way to stay on the right side of that.

Can AI conduct job interviews?

It can run structured first-round interviews. Role-specific questions are drawn from your job criteria, and every candidate gets the same question bank scored against the same rubric, which removes interviewer variability. Final-stage interviews and the hiring decision stay with people.

Do I need to change my HR system to use AI candidate screening?

No. Employment Hero’s recruitment suite is included in the HR plans, or available standalone, connected to the job boards you already post on with no HR migration. 

“The grind wasn’t the job”

We finished by asking Daniella what she’d say to the version of herself who read all 3,994.

“That the grind wasn’t the job. It was the workaround…. The old me didn’t need to click through nearly 4,000 CVs to prove her value. She deserved a tool to handle the sorting, so she could focus on what humans do best, which is making the actual hiring decision.”

Her advice to anyone still doing it the old way is blunt. 

“Stop treating every application as though it deserves the same manual read. Most of a pile of 4,000 won’t be a match, and grinding through it by hand doesn’t make you a better recruiter. It makes you a tired one.”

Try it the way Daniella did

Pick one role. Ideally the one with the worst pile. Get the criteria right, watch the interviews, and see what comes back.

Ready to see it? Book a demo and we’ll walk through it on a role you’re actually hiring for.

About Daniella Angel: Daniella is Talent Acquisition Manager for the Northern Hemisphere at Employment Hero. She has spent nearly fifteen years in talent acquisition across agency and in-house, hiring from entry level through to executive search, and currently leads a team across three time zones. Daniella is a true employment hero.