Candidate experience matters as much as recruiter experience in 2026

A few weeks ago a video went viral on TikTok. A student named Kendiana Colin was being interviewed by an AI bot for a job at StretchLab. The bot glitched and looped the phrase "vertical bar Pilates" for 30 seconds while she sat there. 3.2 million views. The top comment was "Taking the HUMANS out of Human Resources is the worst corporate move ever."

That clip is the most-watched piece of candidate experience commentary of 2026 so far, and the people who run hiring at large companies need to understand what it means. It is not a story about AI being bad. It is a story about a company that bought a tool, did not test it on its candidates, and let the candidates discover the failure in real time.

That is the whole essay in one sentence. Candidate experience in 2026 is not about whether AI is in your funnel. It is about whether you noticed the candidate is on the other side of it.

The financial case, before we get to the moral one

Most blog posts on candidate experience open with empathy. We want to open with money, because the empathy argument has lost three rounds to "we are too busy".

The numbers. 52 percent of candidates have declined an offer after a poor candidate experience (Starred 2025). A positive candidate experience lifts offer acceptance by 38 percent. Average offer acceptance across the market fell to 51 percent in Q2 2025, down from 74 percent two years earlier (HiringThing 2026). 61 percent of US job seekers were ghosted after an interview in 2026, up 9 points since 2024, and 66 percent among historically underrepresented groups (Fortune / iHire). 64 percent of withdrawn candidates withdrew because of poor communication.

Translate that to your funnel. If your offer acceptance is anywhere near the 51 percent market average, half your panel time is currently being spent on people who will not take the job. Most of that loss is not compensation. It is experience. We unpacked the offer acceptance side in a dedicated post, and the candidate experience side is the other half of the same equation.

This is no longer a fluffy metric. It sits on the funnel and the P&L.

Volume hiring breaks candidate experience at a different place than people think

When teams hire at high volume, the assumption is that the top of the funnel is the broken bit. Long application forms. Slow first response. Lousy job descriptions.

The actual data tells a different story. iCIMS 2025 State of Frontline Hiring (2,000 frontline candidates surveyed) found drop-off broke down like this: application 14 percent, scheduling 20 percent, interview 32 percent, onboarding 18 percent. The biggest single drop is not where you applied or how long the form was. It is at the interview itself.

Why. Because the interview is unstructured, the recruiter is overloaded, and one-way video interviews specifically have a 33 percent quit rate (Truffle 2025). The candidate hits a recorded camera, no human on the other side, three behavioural questions, and they close the tab. That is the volume hiring leak.

The fix that actually works is a structured conversational AI screen with three things baked in. Real-time conversation, not one-way video. Citation-backed scoring so the candidate gets feedback even on rejection. Self-serve scheduling around the candidate's calendar. Hilton did this and lifted candidate NPS to 84.9 and cut time-to-class from 6 weeks to 5 days. Unilever did it and saved 50,000 hours of interview time and one million pounds a year while giving every rejected candidate structured feedback. The Unilever case is the most important one in this category because they broke the silent rejection norm at scale. The candidate left the funnel knowing why.

This is what Awesome Hires runs at the top of a volume funnel. Same fairness, same speed, same feedback loop. We covered the operating model of high volume hiring in a separate post, but the candidate experience layer is the half that pays for the rest of the system.

Niche hiring breaks candidate experience for a different reason

The niche hiring case is where most people get the AI conversation wrong, because the loud version of the AI-in-hiring debate is about top of funnel. Senior, specialist, executive hires are a different problem.

The candidate volume per role is low. The interview process is longer. The panel is bigger. Ashby Talent Trends 2025 reports business roles average 11.7 interviews per hire (up 36 percent) and technical roles average 17.6 (up 52 percent). For senior individual contributors and executives the numbers are higher again. The candidate sits through five, ten, sometimes fifteen conversations.

Where does niche candidate experience fall apart. Three places.

Panellists are not prepared. A senior candidate notices instantly when the third interviewer asks the same question the first one already covered. It signals nobody read the previous notes. It signals the candidate is one of many. It signals the company will be like this to work at.

Scheduling is chaos. Senior calendars are a war zone. A two week scheduling delay between rounds tells a senior candidate that hiring them is not a priority. They take the call from someone else.

The panel is taking notes instead of interviewing. Half the panellist is typing, half is listening. The candidate feels it. The signal the panellist captures is shallow, because they were context-switching the whole time. The debrief is unstructured, the rejection is vague, the candidate leaves thinking "those people did not actually want to meet me".

This is what our panel interview product was built to fix. The AI bot sits in the panel call as a silent participant. It captures the full transcript. It generates structured ratings against the rubric. It produces a post-interview summary and Q&A. It feeds signals back into the candidate's profile so the next panellist arrives knowing what was already covered. The human panellists stop taking notes and start interviewing. The candidate gets a more present, prepared, engaged human across the table.

That is the AI in niche hiring that actually improves candidate experience. Not an AI interviewing the candidate. An AI taking the admin off the human so the human can do the thing only a human can do, which is meet another human and decide if they want to work together.

The recruiter side is the upstream cause of the candidate side

Here is the finding from the research that surprised me most. The 61 percent post-interview ghost rate is not a recruiter empathy problem. It is a recruiter capacity problem.

72 percent of recruiters report moderate to severe burnout (Talroo / Pin 2025). 41 percent are considering leaving the profession. 27 percent of TA leaders call their team's workload unmanageable, up from 20 percent the year before. Applications per hire hit 291. The recruiter has 90 candidates in pipeline and 30 minutes a day to communicate.

You cannot fix candidate experience by yelling at recruiters to be more responsive. You fix it by taking 60 percent of their coordination work and giving it to a system that does not get tired. The recruiter ghosting the candidate is not the bad actor in this story. They are the symptom of a funnel that has more applicants than any human can handle.

Recruiter experience and candidate experience are the same metric measured from different sides. Improve one and the other moves with it. That is why every part of the Awesome Hires platform is designed to give recruiters back judgement time and take coordination time off them. The AI screens. The scheduler books. The bot captures the notes. The activity log shows the timeline. The recruiter spends their day on the conversation that needed them, not the calendar that did not.

The metrics that matter

Stop leading with time to hire. It is a vanity metric on its own. Track these five instead.

Candidate NPS by stage. Industry average is 37 (Starred 2025). Hired candidates skew to 80, rejected to minus 5, mid-pipeline to 28. The bigger the gap between hired and rejected, the worse your experience is, because it means people leaving the funnel are leaving angry.

Time to first response. Under 24 hours is the modern bar. Under 4 hours wins. The first response is the first impression.

Withdrawal rate by stage. Where in the funnel candidates leave on their own. Application, screen, panel, offer. Each stage tells you a different story about what is broken.

Post-rejection brand sentiment. What candidates say on Glassdoor and LinkedIn after they left your funnel. This is the metric that compounds into next year's pipeline.

Offer acceptance rate. Already covered in the dedicated post. Lead indicator for everything.

If you track these, you will catch the StretchLab moment before it goes viral.

The legal layer, because it is now a layer

A note on the law, since several teams have asked about this recently.

The EU AI Act classifies hiring AI as high-risk under Annex III, with full deployer obligations from 2 August 2026. Article 86 gives candidates the right to an explanation of the main factors in any AI-influenced decision. Combined with GDPR Article 22, candidates have ten distinct rights when AI is in the loop. Emotion recognition in workplaces and biometric inference of protected traits have been banned since 2 February 2025.

Illinois AIVIA, as amended in 2026, requires explicit written consent before AI analysis of video interviews, 5 business days of pre-notice, and a candidate right to human review by someone with hiring authority within 10 business days. NYC Local Law 144 requires a bias audit and 10 business days of candidate notice. The federal Mobley v Workday class was conditionally certified in May 2025, covering potentially hundreds of millions of applicants.

The throughline is the same in every regime. Disclosure, fairness, real human review on negative decisions, and a candidate right to know why. Build your hiring funnel that way and the law catches up to where you already were. Try to bolt it on later and you are looking at retrofit work in front of a deadline.

The actual point of this post

Candidate experience in 2026 is not a wellness metric. It is the funnel.

Volume hiring breaks at the interview, not the application. Niche hiring breaks at the panel, not the sourcing. Recruiters ghost candidates because they are buried, not because they are bad. AI in the funnel can be the worst thing that happened to candidate experience (StretchLab) or the best (Unilever, Hilton), and the difference is whether the team running it noticed the human on the other side.

We built Awesome Hires on the working assumption that candidate experience and recruiter experience are the same metric. The AI screen gives volume candidates a fair, fast, citation-backed conversation with feedback on rejection. The AI in the panel gives niche candidates a present, prepared, structured human interview. The activity log gives recruiters back the day they were losing to coordination. The intent signal gives the panel a reason to be in the room with the candidates most likely to actually take the job.

The next hiring system you buy in 2026 is the one you are going to live with through both the EU AI Act and your next ghosting story going viral. Make sure it was designed with the candidate on the other side of the screen in mind. That is the part that has changed, and the part that matters.

Sources

• Jabarian, B., & Henkel, L. (2026). "Does AI Beat Humans at Recruiting?" Chicago Booth Review. Field experiment with 70,000 customer service applicants.
• Greenhouse. (2026). Candidate AI Interview Report (n=2,950 across US/UK/IE/DE/AU).
• Starred. (2025). Candidate NPS Benchmarks and Business Impact of Bad Candidate Experience.
• HiringThing. (2026). Job Application Statistics 2026.
• Fortune. (2026). "Candidates ghosted by employers hit three-year high." fortune.com.
• iHire. (2026). "53% of Job Seekers Ghosted by a Potential Employer."
• iCIMS. (2025). State of Frontline Hiring Report (n=2,000).
• Truffle. (2025). Video Interview Statistics.
• Ashby. (2025). Talent Trends Report.
• Talroo. (2025). "The Recruiter Burnout Problem No One Is Tracking."
• Pin. (2026). Recruiter Burnout Prevention Report.
• BestPractice.AI case studies on Unilever and Hilton AI-led hiring funnels.
• NBC News. (2025). "AI job recruiters used by top companies have glitches." nbcnews.com (StretchLab / Apriora viral incident).
• Krebs on Security. (July 2025). "Poor Passwords Tattle on AI Hiring Bot Maker Paradox.ai." krebsonsecurity.com (McHire / 64M applications exposure).
• European Union. (2024). EU Artificial Intelligence Act, Article 86 and Annex III. Official text.
• Illinois General Assembly. HB 3773 / AIVIA (2026). Official statute.
• New York City Department of Consumer and Worker Protection. NYC Local Law 144. Official page.
• Mobley v. Workday, Inc. (N.D. Cal., 2025). ADEA conditional certification, May 2025.
• UK ICO. (May 2026). AI in recruitment guidance. ico.org.uk.