The Candidate Your AI Just Rejected Is Also a Customer
A woman named Louise applied for a job at Virgin Media.
The interview was a mess, and not because of her. A grumpy receptionist. An interviewer who stepped out to take a phone call, came back ten minutes later, said he'd heard everything he needed to hear, and rejected her on the spot.
Louise was a Virgin Media customer. She went home, cancelled her subscription, switched to a competitor, and told her sister to do the same.
That was one candidate. Virgin's head of resourcing, Graeme Johnson, started asking the expensive question: how many more Louises are out there?
His team pulled the data. In 2014 they surveyed the roughly 123,000 people Virgin had rejected that year. 18% were existing customers. About 7,500 cancelled within four weeks of being rejected. At £50 a month, that was roughly £4.4 million a year in lost revenue, almost exactly what Virgin was spending on recruitment in the first place.
They had built a machine that spent millions attracting people, then quietly cost millions in how it treated the ones it turned down.
Yesterday I wrote about the other half of this story... how AI screening doesn't strip bias out of hiring, it learns yours and scales it, quietly shaping who gets screened in. This post is about everyone it screens out, and how they feel being judged by a machine in the first place.
That was 2014. And a human did it.
Here's what's changed. That whole candidate experience, the part Virgin learned to guard so carefully, has since been handed to actual machines.
AI now sits across the entire funnel: sourcing, résumé screening, scheduling, one-way video interviews, scoring, and the rejection email itself. HireVue alone ran more than 20 million one-way video interviews in a single quarter of 2024. By some estimates around 90% of employers now use automated screening somewhere in the process. For a growing share of applicants, the first "person" they meet at your company is not a person.
So the question Johnson asked by accident is now a question every business is answering on purpose, thousands of times a day, through software. And candidates have opinions about it.
How job seekers actually feel about it
This is the part that should worry you, because the data got specific in the last year.
In a 2026 survey of nearly 3,000 job seekers, 63% had already been interviewed by an AI, up 13 points in just six months. 70% of them were never told AI was involved. And roughly half have either walked away from a hiring process because it used an AI interview, or say they would.
The mood underneath that is uneasy. Two-thirds say they feel uncomfortable with AI-led hiring. Only about a quarter trust AI to evaluate them fairly. In a separate 2026 study, nearly half of candidates said they believe these tools are biased against their age, race, gender, or background, and that number climbs for neurodivergent applicants. Half reported being rejected without a single word of explanation. One described being told by the software to "better align with an ideal personality profile," with no hint of what that meant.
That is Louise's experience, rebuilt in code and shipped to everyone at once.
The important nuance, before you rip it all out
This is not a torches-and-pitchforks story, and pretending it is would be wrong.
Most candidates are not anti-AI. Other 2026 surveys find large majorities are fine with, even grateful for, AI that gives them a fast answer, flexible scheduling, and a consistent process. Only about one in five actually want less AI in hiring. Plenty of people would rather do a screening on their own time at 9pm than play phone tag with a recruiter for a week.
So what are they rejecting? Not the technology. The opacity. The coldness. The silence. Being judged by something they can't see, can't question, and were never told was in the room. Candidates aren't asking you to fire the robot. They're asking you to tell them it's there and treat them like a person anyway.
Why this is a revenue problem, not an HR one
Every one of those walk-aways and black-hole rejections is a modern Louise. The difference is scale.
In 2014 Virgin created detractors one rude interviewer at a time. Today you can create them by the thousand, automatically, and mostly without knowing it happened. The candidate who abandons your one-way video interview, or gets the 6:04pm auto-rejection with no explanation, doesn't file a complaint. They just quietly decide how they feel about your brand, and then they act on it. Around 4 in 10 people with a bad candidate experience say they'll stop buying from the company. Roughly 7 in 10 tell someone.
The cost Virgin Media measured by hand is now being generated by your software, at a volume no human team could ever have matched.
Why sales leaders should care most
For a consumer brand the applicant-customer overlap is obvious. For sales leaders it's sharper.
Your sales candidates come from your own industry. They sell for your competitors and partners. Some sell to, or buy from, the exact accounts you are chasing. Your candidate pool and your market are the same room.
Now put a cold, opaque AV interview and a silent auto-rejection between you and that room. A strong closer reads that process the way a buyer reads a bad first sales call, because that's exactly what it is: your company's first impression, automated. The candidates with the most options, the ones you most want, are the first to walk. And they talk.
What doing it right looks like
Here's the thing candidates are actually asking for, and it isn't a faster auto-reply. They want to know a real person is involved, that a human made the call, and that they can reach one if they need to. In the 2026 Greenhouse survey, 38% said plainly what they want is confirmation that a person reviews the AI's output before any decision is made. Responding quickly is the floor. Visible, reachable human involvement is the bar. Five steps to build it into a high-volume, AI-heavy process.
Make the human involvement visible, not just present. It isn't enough to have a recruiter somewhere in the building. Tell candidates a named person reviews the AI's output before any decision, and say where in the process that happens. Disclose the AI too, and what it measures, because 70% of job seekers were never told it was there. But the reassurance they're really after is that a human is still in the loop.
Keep the real decisions human, and make sure the human actually decides. Roughly 7 in 10 people oppose letting a machine make the final hiring call. Let AI clear volume and surface a shortlist, but a person owns the decision and delivers it, especially the rejection. And that person has to genuinely review the AI's calls, including a sample of who it screened out, not just approve the list. Half of candidates were rejected in the past year without a single word from a human, and most assumed a machine did it. A rejection signed by a named recruiter is the minimum proof that someone actually saw them.
Give every candidate a way to reach a human. Not a chatbot that loops back on itself... a named contact, and an explicit option to ask a question or request a human review of an AI decision. One 2026 survey put the share who want that option at nearly three-quarters. Most will never use it. The point is that it exists and they know it does. Regulators are heading the same way: frameworks in New York and the EU already require offering a human alternative.
Spend your human hours on real conversation, where it changes the outcome. You can't hand-hold four thousand applicants and you shouldn't try. Put actual conversation at the moments that carry weight: a live human interview before any final decision, and offers and late-stage rejections delivered by a person, not a portal. It isn't only kinder. Companies that pair AI screening with human-led final interviews report faster hiring and better first-year retention. The blend beats either extreme.
Measure it against revenue. Virgin Media only found its £4.4 million because someone cross-referenced rejected candidates against cancelled accounts. Run your own version: track where candidates drop off or walk away, and match your reject pile against your customer list and target accounts. The number exists whether you look at it or not.
The part I actually tell clients
Most companies still think of the screening stack as a back-office efficiency play. It isn't. It's a customer-facing system now. The AI reading the résumés and running the interviews is making a first impression on a big slice of your market, every single day, and it's doing it in your name.
In sales, where reputation travels fast and your candidates and customers drink at the same bar, that first impression is part of your go-to-market whether you designed it that way or not.
Virgin Media spent millions learning that a rejection is a brand moment. The tools got cheaper since then. The lesson didn't change.
Speed and humanity aren't a trade. You can run a fast, automated, modern hiring process and still treat the people in it like they might one day sign your invoice. The companies that figure that out won't just hire better. They'll stop quietly bleeding customers they never knew they had.
In short
A rejected candidate is often also a customer or someone who talks to your buyers, and Virgin Media traced roughly £4.4 million a year in lost revenue to how it treated rejected applicants. Today AI runs much of that experience, from résumé screening to one-way video interviews to the rejection itself. Most job seekers have now been interviewed by AI, most weren't told, and about half have walked away or would. Candidates aren't anti-AI; they value the speed but reject the opacity and silence. What they actually want is visible human involvement: to know a person reviews the AI before it decides, that a human owns the call, and that they can reach one. The fix is to make that human involvement visible and reachable, keep the real decisions and the rejection with a person who genuinely reviews the AI, put human conversation at the moments that matter, and measure candidate experience against revenue.
Key takeaways
A rejected candidate is frequently also a customer or a contact in your market; Virgin Media traced about £4.4 million a year in lost revenue to how it treated the people it rejected.
AI now runs much of the hiring funnel, including screening, one-way video interviews, scoring, and rejections; HireVue alone ran more than 20 million one-way interviews in a single quarter of 2024, and most employers use some automated screening.
In a 2026 survey of nearly 3,000 job seekers, 63% had been interviewed by AI (up 13 points in six months) and 70% weren't told it was involved.
Roughly half of candidates have walked away from a process because of an AI interview or say they would; about half feel these tools are biased against them; only about a quarter trust AI to judge them fairly.
Candidates aren't anti-AI. Most value the speed and flexibility and only about one in five want less of it. They reject opacity, coldness, and black-hole rejections, not the technology.
What candidates want most is visible human involvement: 38% specifically want confirmation a person reviews the AI before it decides, and roughly 7 in 10 oppose a machine making the final call. The fix is to make human involvement visible and reachable, keep the real decisions with a person who genuinely reviews the AI, put human conversation at the moments that matter, and measure candidate experience against revenue.
Frequently asked questions
Do job seekers mind AI screening and interviews?
It's mixed, and the split matters. Most candidates are not opposed to AI itself; large 2026 surveys find majorities value the speed, flexibility, and consistency it can bring, and only about one in five want less of it. What they object to is opacity and coldness: being screened or interviewed by AI without being told, getting no feedback, and being rejected into silence.
How many job seekers have been interviewed by AI?
In a 2026 survey of nearly 3,000 job seekers, 63% said they had been interviewed by an AI, up 13 percentage points in six months. Notably, 70% of them were not told AI was involved.
Does AI hurt candidate experience?
It can, when it removes transparency and human contact. Roughly half of candidates in 2026 research have walked away from a hiring process because of an AI interview or say they would, about half believe the tools are biased against them, and only about a quarter trust AI to evaluate them fairly. Because rejected candidates are often customers, that damage carries a revenue cost, as Virgin Media's £4.4 million example showed.
What does this mean for sales hiring specifically?
Sales candidates usually come from your own industry and often deal with the same buyers you are targeting. A cold, opaque AI process is effectively a bad first sales call to people inside your market, and the strongest candidates, who have the most options, leave first and talk about
How do you use AI in hiring without driving candidates away?
Make the human involvement visible: tell candidates a named person reviews the AI's output before any decision. Keep the real decisions, especially the rejection, owned by a person who actually reviews the AI rather than rubber-stamping it. Give every candidate a way to reach a human and request a human review. Spend your people's time on real conversation at the moments that matter, like a live final interview. Then measure candidate experience against customer and revenue data. Responding quickly matters, but it's the floor, not the point.
References
Clutch. (2026, July 23). The state of AI hiring in 2026. https://clutch.co/resources/state-ai-hiring
Enhancv. (2026, June 8). AI hiring in 2026: Half of job seekers were rejected without a word. https://enhancv.com/blog/ai-hiring-statistics/
Fast Company. (2016, April 5). How your company's lame hiring process is losing you customers. https://www.fastcompany.com/3058571/how-your-companys-lame-hiring-process-is-losing-you-customers
Greenhouse. (2026, May 1). 63% of job seekers have faced an AI interview. Most haven't had a good one yet [2026 Candidate AI Interview Report, n=2,950]. https://www.greenhouse.com/newsroom/63-of-job-seekers-have-faced-an-ai-interview-most-havent-had-a-good-one-yet
HiringBranch. (2026, July 16). Artificial intelligence resume screening. https://www.hiringbranch.com/blog/artificial-intelligence-resume-screening
LinkedIn Talent Blog. (2017, March 15). Bad candidate experience cost Virgin Media $5M annually, and how they turned that around. https://www.linkedin.com/business/talent/blog/talent-acquisition/bad-candidate-experience-cost-virgin-media-5m-annually-and-how-they-turned-that-around
MSH. (2026). Current recruitment and hiring trends in 2026. https://www.talentmsh.com/insights/hiring-recruiting-trends-statistics
SelectSoftwareReviews. (2026, May 25). Latest AI recruiting statistics: What the data says about hiring in 2026. https://www.selectsoftwarereviews.com/blog/ai-recruiting-statistics
TalentAdore. (n.d.). How candidate experience is causing a huge impact on businesses. https://www.talentadore.com/resources/candidate-experience-causes-huge-impact-businesses
