Job recruitment didn’t always work this way. Employers once read applications, and agencies actively found roles for their candidates, matching people to opportunities and advocating for them. Cover letters were read, not assumed to be generic filler.
Today both employers and agencies often rely on automated scoring, keyword filters, and rigid matching rules. Agencies no longer search for jobs on behalf of applicants; instead they attempt to slot people into the vacancies they already hold, discarding anyone who doesn’t fit the preset criteria. Tailored, job‑specific cover letters are frequently not opened at all, wrongly presumed to say nothing more than “Please find my CV attached.” In this system, most applications never reach a human being, and experience or potential can disappear long before anyone has the chance to see it.
AI screens and rejects
Now that automated hiring systems are embedded in recruitment, the process no longer simply filters candidates, it defines their visibility. Applications are parsed and filtered through applicant tracking systems (ATS), then scored and ranked, using AI-driven models, before a recruiter ever engages with them. Employers and agencies depend on these platforms to determine who is even worth reviewing. Systems such as Be Applied, Greenhouse, Workday, SmartRecruiters and iCIMS are no longer just administrative tools; they form the decision-making layer that determines who is seen and who is excluded.
These recruitment platforms are used to manage and process applications more efficiently. They can automatically screen CVs, score short answer responses, and rank candidates based on specified skills, experience or other job-relevant factors. Some systems anonymise applications, shuffle or segment responses, and use structured tests to assess problem solving, analytical or numerical abilities. The goal is to prioritise candidates for human review by identifying those who meet predefined criteria, while reducing the time recruiters spend manually sifting through applications.
The human touch and interpretation is gone
When you write a response expecting a human to read it, you naturally include context, nuance, reasoning, and personality. You may explain why you handled a situation a certain way, give examples, reflect on outcomes, or demonstrate your thought process. You also use phrasing that conveys tone, empathy, judgment, or creativity, things a human can intuitively interpret.
An AI-driven system, however, typically does not read for nuance. It scores your answers mechanically, focusing on:
- Keywords and phrasing that match the job description or pre-set skills
- Structured data points, like years of experience, qualifications, or numerical answers
- Performance on timed or multiple-choice assessments
What may get missed:
- The narrative of your experience and why it matters
- Insight into your judgment, priorities, or problem-solving approach
- Personality, voice, and evidence of soft skills
- Subtle connections between your past experience and the specific role
In other words, AI reads the signals it can measure, not the story you are telling. Even carefully thought-out, human-focused answers can be invisible if they do not hit the measurable markers the system expects. This is why your experiences and reflections might never reach a human reviewer unless your score is high enough.
Daisy’s story of wanting to work
These changes aren’t abstract. They affect real people every day. I interviewed a woman in her late fifties who wishes to remain anonymous, so I will call her Daisy. She is 56 years old and has over 40 years’ experience as an executive assistant, office manager, and administrator. She had maintained steady employment until her redundancy in 2024. After that, things started to go wrong.
When Daisy was made redundant in September 2024 she turned job-hunting into a full-time role, working from 8am until 4pm. In the first month alone, she applied for over 300 positions. Since then, the total has climbed to roughly 1,000 applications across multiple job boards and direct company submissions.
Despite being qualified, proactive, and working harder than ever, she secured only five interviews. Yes, only five.
Here is the disturbing truth: Daisy is no longer being rejected by people. She is being erased by systems that never even see her.
While many professionals are struggling in this economic climate, her story highlights a darker trend in UK recruitment: the rise of “black box” hiring, where human experience is filtered out by rigid algorithms and automated tests before a single recruiter ever reads your CV. Experience, loyalty, and decades of work now carry little weight in a system that prioritises speed over substance.
Irony meets reality
Daisy recently applied to a not-for-profit organisation dedicated to fairness in technology. Despite adequate qualifications, the application was processed entirely by an algorithm. Only candidates with a specific score threshold were to be contacted. No human read the application. The irony of a “fairness” organisation using an unfair system is stark., More concerning still was a line in the job description which read:
“Xx Not for Profit Company does not use AI in its recruitment processes, except to detect applications for AI use. As a tech-justice organisation, we ask the same of our candidates.”
Yet, the platform that processed her application was hosted by Be Applied, a system that does exactly that.
The speed test barrier
Another role required an intensive application form, five expansive written questions, and a timed skills test. Her thoughtful responses were never seen by a person; only a computer which scored them to determine progression. She was rejected not necessarily for lack of ability, but because she didn’t meet an arbitrary digital metric and was not in the top 1 or 2 percentile.
Intensive interviews
One company required three interviews for a remote role, spread over a month: first with the manager, then with the manager and their superior, and finally with the Chief Executive. The final interview was to take place in London on a single specified date, which she could not attend., She advised the company of this, and offered them alternative dates, but none were accepted. Instead she was offered another Teams interview.
At each interview (apart from the first), she was given a “task” to discuss:
- Fundraising ideas specific to the charity;
- Add another word to the current list of values.
Here lies the striking irony. In every interview Daisy had ever experienced over her 40-year career, she had been asked some variation of “Tell us a little bit about you?” It is the universal human opener that allows candidates to weave their narrative, share their motivations, and connect as people. Yet across these three interviews, not once was she asked anything about herself. They bypassed her story entirely, treating her instead as a vessel for specific data points to fill into their “values” list or fundraising buckets.
The process was so focused on extracting the right answer that it forgot to interview the person. Daisy was ultimately unsuccessful. When she asked for feedback, the response revealed exactly where the disconnect lay:
Positives:
- You were well prepared in regards to researching the Trust
- You were open and personable
- Your answers were considered, especially with regards to successful company culture and conflict.
Negatives:
- We would have liked you to have expanded on your answers, providing more detail and linking back to the skills which were listed in the person specification
- It would have been beneficial to have met you in person at the suggested interview location
- Another candidate had stronger experience in charity fundraising environments
In effect, she was disadvantaged for missing a single specific date, criticised for answers that were human-focused rather than formulaic, and ultimately passed over for a more “plug-and-play” candidate.
The human cost
Even when she did receive job offers, the human element often failed her. One role involved workplace bullying that shattered her confidence; another left her uninformed, unpaid and ghosted by a senior director; and a third company suspended hiring mid-process after weeks of silence. In one reception role, staff feedback was delivered via email with an attached audio file that could not be accessed until after a shift had ended, an impersonal and impracticable solution.
The core issue
Throughout this process, it has become clear that we are losing the “human check” in recruitment. Employers now often rely on brief aptitude tests and keyword filters to decide who even reaches an interview stage. With hundreds of applications being received per role, the days when a strong CV could catch a hiring manager’s eye appear to be fading.
In an era where we are told brevity in your CV is the key, no more than two pages long, we are effectively removing keywords from our CVs. When you condense experience by that amount, your application effectively becomes ‘applicant light’. Although ATS don’t penalise length, human recruiters do. So a 4-page document is seen as cumbersome and unfocused. The standard for even senior roles is 1-2 pages, with anything over 4 pages being viewed as excessive. Candidates are therefore forced to choose between being visible to machines or readable to humans, and either choice carries a cost.
With 1,000 applications yielding only five interviews, the current system is clearly failing experienced workers. Daisy’s story is not unique, thousands of professionals are becoming invisible to systems designed to identify talent, rejected by algorithms rather than people. The wider trend is, luckily, less desolate, with around 2-3% of applications resulting in interviews – that’s around 42 applications to one interview. However, the underlying problem remains: experience and skill are often invisible to the system.
What do we do now?
In a world where algorithms increasingly pre-filter applications, candidates must adapt without losing their human strengths. A few practical approaches are:
- Understand the systems. Platforms prioritise keywords, skills, and structured responses. Tailoring CVs and answers to match the role’s required skills and phrasing can help your application reach a human reviewer.
- Use concise, measurable examples. Even when writing for a human, make sure key skills, achievements, and outcomes are clearly stated in ways the AI can detect. Numbers, tools, qualifications, and specific skills matter more to the system than subtle nuance.
- Network strategically. Many roles still involve referrals, LinkedIn connections, or direct introductions. Human connections can bypass or supplement algorithmic filtering.
- Track your applications and adapt. If a pattern emerges that certain types of wording or formats get more responses, adjust future applications accordingly.
- Keep the human story alive. Even if AI reads the first round, your human insight, judgment, and personality are what will convince a hiring manager in an interview. Do not stop crafting thoughtful narratives; just pair them with AI-friendly structure so they actually get seen.
It is about balancing human strengths with algorithmic realities, ensuring the system recognises you long enough for a person to meet you.
The human truth
When I asked Daisy what she had learned from this process, her answer was stark:
“I now spend over an hour sculpting my CV to highlight the specific skills each role requires, and write personalised cover letters that explain why I want to work for them. But I have learned a hard truth: employment agencies do not read these words, not a single word. Creating one for them achieves nothing. So, if I create a cover letter at all, it is just generic filler.”
In other words, she frets over every comma and keyword, ensuring the algorithm sees her, only to realise that once she passes the gate, the human element has already been bypassed or ignored anyway. She has adapted perfectly to the machine’s demands, yet the machine remains indifferent.
I would love to tell you this tale has a happy ending, but it doesn’t. Daisy is still unemployed, applying for roles for which she is fully qualified, and receiving rejection emails, or more often than not, absolute silence.
She has optimised for the system. The system has ignored her.
AI 995 – Daisy 5.
Friends of Bylines Network
There has never been a greater need for grassroots journalism that investigates the stories that really matter, holds power to account and champions the voices of everyday citizens. We are proudly powered by volunteers but what we do isn’t free.
STAND WITH US for independent, citizen-led journalism that makes democracy stronger, and you will even get some exclusive benefits.
