Participants who have watched employers recruit describe the people and decisions they saw. They also say which parts of the process they did not see.

"At the foundation, managers read the forms. Two of them disagreed about a candidate while I was there. One liked the community experience and the other wasn't happy with the reporting example. They went back through it. I don't know what happened before it reached them."

— P14 · Business management · Q8

Here, two managers disagree and read the application again. P14 saw that discussion, but did not see how the application reached them.

"My placement manager read CVs and gave people a practical task. I saw that part. I didn't see what HR did before him, so I can't tell you the whole process."

— P18 · Cybersecurity · Q8

"An engineer sees it eventually. I saw the managers talking about candidates on placement. I didn't see the bit before that. Some sort of check by HR, I'd imagine."

— P02 · Mechanical engineering · Q8

"HR first, then the home manager I think. I've sat next to managers reading the forms. A small home might not have the HR bit."

— P07 · Health and care management, online · Q8

One participant reports a workplace application count that a manager showed her:

"We got 350 for reception. My manager showed me."

— P11 · Business and tourism management · Q11

Participants without that exposure describe an end-to-end pipeline. Asked where the picture came from, they name media sources rather than people or workplaces.

"ATS first. It scans the keywords, then someone in HR looks. That's what I thought anyway." … "Videos. Loads of them say it. I haven't actually seen anyone recruit, so I can't tell you for sure."

— P16 · Business management · Q8

P16 believes this because so many videos repeat it. P06 gives a more technical explanation, based on a podcast and something he saw on GitHub. Neither has seen an employer use the process they describe:

"It hits an ATS, like Workday or whatever they use. It gets scored on keyword match and criteria, and ranked. For the big firms a recruiter might only look at the top of the ranked list, so loads never get human eyes at all." … "The technology, mainly. I know what it can do, I listened to a podcast once and saw a commit on github from a guy that works there."

— P06 · Computing, online · Q8

"I imagine they check the grades and university first. Then keywords. I don't know, that's what all the advice makes you think." … "Mostly forums and videos, honestly. And being rejected. You start wondering what they didn't like when they don't tell you."

— P17 · Law · Q8

When employers give no reason for a rejection, P17 turns to online advice to try to understand what went wrong.

"They probably test people first. There are too many CVs to read. Some bot must sort them out, I'd think."

— P09 · Business and computing · Q8

The account students absorb describes a large-employer graduate scheme. Many participants are not entering one. The accounts suggest a tension between a generic screening model and the different recruitment routes participants describe. They do not establish what software any employer uses.

"I think the ward manager or recruitment team reads the form and scores it against the person spec. That's what we've been told on placement."

"It doesn't really feel like a graduate job search to me. More like finishing placement and qualifying."

— P05 · Adult nursing · Q8, Q19

"A lot of it is who knows you. The broadcaster schemes have forms and tests, but I've seen other jobs get filled before anyone even puts them online." … "Somebody knows somebody and the job's gone. If you don't know anyone yet, where do you start?"

— P12 · Film, TV and digital production · Q8

"At a small brand it could just be the founder opening your email and looking at the portfolio. That's closer to what I saw on the internship."

— P10 · Fashion marketing · Q8

"A hundred? For the big-company one. Care's different, we struggle to get people."

— P07 · Health and care management, online · Q11

A labour-shortage sector, described by someone working in it, while the surrounding discourse describes scarcity.

"The graduate schemes tell you the stages. Forms, tests, interviews, assessment centre. I've been through some of those. At work we use an agency and the manager interviews people. I don't think there's one process for all of them."

— P20 · Business management · Q8

That last account explicitly distinguishes recruitment routes and resists generalising from one to all others.

Participants describe several bases for judgment, ranging from external references to familiarity with their own experience and craft.

Against their own experience. Participants use what they know about their work to spot invented achievements and check technical claims.

"If it's made up something about my job, I'll know. I check it against my figures and what they've asked for."

— P03 · Business and marketing, online · Q12

"The technical stuff I can check. If something looks odd, I'll look at the docs or try it myself."

— P06 · Computing, online · Q12

"It got an event ID wrong, so I went and checked that. Still useful practice."

— P18 · Cybersecurity · Q3

Against how they actually speak.

"For language, I can judge it cuz I couldn’t write like that. I also read the output aloud to see if it sounds like something I could actually say."

— P08 · Criminology and law, online · Q12

Against craft judgement. One participant describes a route the others do not have — she can see exactly what a tool changed.

"With the editing software I can see what a feature has changed. It's easier to judge that than a whole generated idea."

— P12 · Film, TV and digital production · Q4

One participant combines several forms of checking:

"Company information, their website. Technical things, course notes or someone qualified. With wording I check that my manager would recognise the job I'm describing. If it makes me responsible for everything, something's gone wrong."

— P20 · Business management · Q12

The same participants describe a hard stop when the claim concerns recruitment itself.

"The recruitment stuff is harder. I don't know what settings they've got."

— P06 · Computing, online · Q12

"Technical stuff, test it or check the documentation. Recruitment advice is harder. I ask people who've hired analysts. I don't put much weight on a CV score."

— P18 · Cybersecurity · Q12

"When it starts telling me about ATS scores, I don't know how to check that."

— P03 · Business and marketing, online · Q12

"I don't really know. If it sounds generic I don't like it."

— P04 · History · Q12

"If you get the job maybe? I’m not sure."

— P11 · Business and tourism management · Q12

That last answer is worth sitting with. The only test she can imagine is the outcome — which arrives long after the decision it was meant to inform, if it arrives at all.

When an external standard is difficult to access, participants describe other signals that feel informative. An opaque score, institutional endorsement, agreement between tools and AI reviewing its own output raise different questions about the basis for trust.

A score.

"I like seeing the score go up. It makes me feel I've improved it, even if I can't really tell from reading it."

— P19 · Business management · Q4

Institutional branding.

"The university logo made it feel approved I guess so I treated the score more seriously. I don't know whether it uses the same sort of model as everything else."

— P09 · Business and computing · Q12

Agreement between tools.

"I compare the answers. If two tools say the same thing, I tend to believe it more. I know that might not mean much."

— P16 · Business management · Q12

Asking the machine to review itself.

"I know I'm asking the same thing that wrote it, but I still do it."

— P19 · Business management · Q12

The awareness is not the missing piece. Some participants recognise the weakness of a signal while continuing to find it reassuring. Alternatives may exist but feel less accessible or immediate.

Keyword insertion continues alongside expressed scepticism; it is described as something concrete the participant can do.

"Not that certain, actually. I know they use the systems. I don't know the settings. I still put the keywords in because well at least that's something I can do."

— P03 · Business and marketing, online · Q8

"Matching what they ask for. Experience is probably the main thing …" … "I still think the keywords matter."

— P16 · Business management · Q9

Without a standard, there is no signal that a draft is finished. Several participants describe revision expanding to fill the time available.

"I couldn't choose. One gave it a better score than the other and I was trying to work out why. It was only about three hundred words. I spent most of the evening on it."

— P17 · Law · Q3

"Can it make me stop editing? That's what I'd want. Give me a couple of versions and make me choose. I'd probably open another chat, actually, so I don't know if that would work."

— P17 · Law · Q18

"Like how do you know when enough is enough cuz you can really spiral outta control. I can spend hours trying different versions until I no longer know which one is clearer."

— P09 · Business and computing · Q14

"Bits of it. Then I asked it to compare the two versions. By the end it was close to what I'd written in the first place. I do that quite a lot."

— P04 · History · Q3

"I'll feel better for ten minutes, then it gives me five alternatives and I wonder what was wrong with the first one."

— P01 · Psychology · Q19

In this account, a deadline or another person provides a stopping point:

"The deadline. Or my sister saying enough now. Usually the deadline."

— P17 · Law · Q14

The same reasoning appears across completely different fields, in participants' own words.

"Not when I picture the interview. They'll ask what I actually did, and then I'm explaining why the CV doesn't mean what it says."

— P18 · Cybersecurity · Q13

"It once added a load of skills to my CV I'd never claimed, like being proficient in data analysis software, and I deleted them because I'd have been caught out instantly in an interview."

— P01 · Psychology · Q13

"It'll make it sound like I transformed a society when I was just a member. I take that out. If they asked me about it, I wouldn't have anything to say."

— P04 · History · Q13

"They might. I'd be more worried about not being able to explain the answer. If they ask about my work, I need to know exactly what I've claimed."

— P20 · Business management · Q15

That answer was given to a question about whether employers can detect AI writing. He ranks defensibility above detection without being prompted to.

The catches themselves are precise:

"It said I advised senior stakeholders on regulatory risk. I help in a shop."

— P17 · Law · Q13

"It called a phishing exercise red-team experience. And monitoring became threat hunting. They sound close if you don't do the work, but they're different things."

— P18 · Cybersecurity · Q13

"It also tried to turn attendance going up into revenue, which we hadn't measured."

— P14 · Business management · Q13

"It also called the documentary award-winning which was hilarious cuz it only screened at our course showcase …"

— P12 · Film, TV and digital production · Q3

"it made my supporting statement sound really formal and business-y, I rejected most of it. It didn't sound like a caring nurse, it sounded like a manager."

— P05 · Adult nursing · Q13

"It suggested chartered accountant for my headline, as an ambition. I haven't qualified. I kept student and accounts assistant. I don't want someone reading it quickly and getting the wrong idea."

— P20 · Business management · Q13

Participants identify three conditions under which it fails, and they name them themselves.

Tiredness.

"When I'm exhausted I don't check as well. I know I should. Sometimes I just want it sent."

— P11 · Business and tourism management · Q14

Volume.

"Definitely. The first one gets checked properly. By the last one I'm just trying to get it sent."

— P19 · Business management · Q2

Familiar vocabulary. The subtlest of the three, and the one participants are least able to guard against:

"Not straight away. It was using words I recognised, so I read past it. Then I thought about them asking me how the system works and realised I'd made it sound like my responsibility."

— P16 · Business management · Q3

"I took it out, but some of the smaller stuff is harder to notice."

— P16 · Business management · Q13

And when it fails, nothing tells anyone:

"One application said I managed a regional promotion. I managed it in my store and sent feedback to the region. I noticed after submitting. I didn't get an interview, so nothing happened but I’d hate to be caught out on these types of lies I’d be mortified."

— P03 · Business and marketing, online · Q13

One exchange shows a stated practice giving way under a specific question:

"I think so. I'd have to open it. There are so many versions. I normally say I check the numbers, but now you're asking about a specific one I'm less sure."

— P19 · Business management · Q13

Some participants use AI to improve their written English. They worry that an interviewer will hear them speak differently and mistake that help for dishonesty.

"I worry that the application will sound more fluent than I am when speaking, and the interviewer will think I was dishonest. I try to keep the application to be clear, not to pretend that English is my first language."

— P08 · Criminology and law, online · Q14

P08 holds back on polishing his English because he does not want an interviewer to think he lied.

"When they meet you and your English is different, probably. But people get help with English anyway. I don't think that should mean you can't have the job."

— P11 · Business and tourism management · Q15

"I need the steps first. A big block of text is difficult with dyslexia. So I get it to pull out what they're asking for, then do one bit of my CV at a time. I listen to the changes as well."

— P09 · Business and computing · Q5

"And when it changes my English, explain why. Otherwise I just copy the sentence."

— P08 · Criminology and law, online · Q18

P08 wants to understand the changes so he can learn from them, rather than just copy the new sentence.

Accounts from participants with professional experience connect confidentiality principles to AI use, and describe recognising risks beyond simply removing a name.

"I never paste names, details of residents or an incident report into it they really drilled that into us."

— P07 · Health and care management, online · Q14

"Confidentiality. Taking a name out doesn't always make the story anonymous. I nearly typed a work example in once and then thought, no, there's too much in that."

— P15 · Business management · Q14

"I use made-up events for practice, not client logs."

— P18 · Cybersecurity · Q14

"I take the private details out first. It's not worth putting somebody's information in there for a better paragraph."

— P05 · Adult nursing · Q14

These accounts suggest a principle carrying across contexts. They arose in response to questions about concerns and risks; they do not establish that every participant learned it in the same way.

"I ask AI first now. I used to ask people more. It's easier when it's late and I don't want to bother anyone. Sometimes I think I use that as an excuse, though."

— P17 · Law · Q17

She establishes the absence earlier in the same interview: "I haven't got anyone in the family who's done law, so there's a lot I'm trying to find out."

"Most of the writing goes through it now. I used to ask friends more. You can ask ChatGPT ten times and it doesn't get annoyed, which is probably why I don't stop."

— P19 · Business management · Q17

Set against participants with people to ask:

"I'd ask the placement team if the CV was any good. Or my friend who's an engineer. AI's fine for a sentence."

— P02 · Mechanical engineering · Q12

"If it points to something I've missed, I can check that. The general advice, I'm less sure. I'll send it to my old supervisor if it's important. She doesn't write much back, but it's usually useful."

— P14 · Business management · Q12

"The person spec, mostly. And I'd ask my mentor. I trust her on what I should say about placement."

— P05 · Adult nursing · Q12

"I use it to get ready. The contacts are still where I hear about jobs."

— P10 · Fashion marketing · Q17

Alongside practical access and perceived relevance, participants describe a social cost to asking for help.

"Partly timing and partly embarrassment, if I'm honest. I feel I should arrive with a finished plan."

— P01 · Psychology · Q16

"I don't really want to explain how many applications I've sent." … "They might say I've been doing it wrong. Which could be useful, I know. I still don't want to hear it."

— P19 · Business management · Q16

She names the feedback as valuable and avoids it anyway.

"The appointment was good. She suggested local firms I hadn't been looking at and got me to talk through the shop example. I should book again. It feels awkward coming back with the same problems."

— P17 · Law · Q16

"It's my first place for questions now. You can ask something really basic and it doesn't matter. I find it harder to ask a lecturer when I feel I should already know the answer."

— P16 · Business management · Q17

"It feels close enough for remembering examples but not for pressure or judgement. A person is obviously better but realistically you’re not gonna have someone with enough patience for you to rehearse these things."

— P12 · Film, TV and digital production · Q6

In these accounts, patience and the absence of judgment are advantages even where confidence in advice quality is limited.

Where participants describe careers support working, it did something specific: it changed how they understood the problem, rather than improving the text.

"She asked if the jobs would actually get me where I wanted. I'd been asking ChatGPT to fix the applications, not really thinking about that."

— P03 · Business and marketing, online · Q16

"The recording showed I wasn't listening when someone raised a risk. I acknowledged it and carried on with my list. I hadn't noticed myself doing that." … "I didn't spot the listening thing myself, and an AI-written answer about teamwork wouldn't have shown me that."

— P20 · Business management · Q6, Q17

"There was a shot I really liked, but another student designed and lit it. I was assisting. The tutor asked what it showed about my work. I didn't like taking it out, but I could see why."

— P12 · Film, TV and digital production · Q16

"I had a list of duties. At the workshop they got me to use the night handover instead. We'd had the same problem coming up and I got people to use a clearer checklist. I hadn't thought of putting that on the CV."

— P11 · Business and tourism management · Q16

"Stop writing like a supporter. Explain what I could do for them. That was before the placement application, and I still catch myself doing it."

— P14 · Business management · Q16

"She showed us CVs from people who'd been hired. And what the job was actually like, chasing samples, changing product copy. That was useful."

— P10 · Fashion marketing · Q16

These examples include reframing, noticing behaviour and selecting relevant experience, rather than wording alone. Some participants contrast that experience with an AI tool’s tendency to agree:

"ChatGPT tends to go along with whatever I say I want to do there’s never any pushback which is not great."

— P03 · Business and marketing, online · Q17

"Sometimes I feel the AI is too positive and is only trying to fulfill my request to what it considers to be the best result …"

— P21 · Business management · Q13

Asked to design ideal support, participants split. One group asks for judgement.

"Help me work out which jobs I can realistically do. Before it writes the application. I can get a nice statement already, that's not the problem."

— P16 · Business management · Q18

"I'd want it to tell me if something's bad. Not just say I'm great. And then actually help me fix it."

— P01 · Psychology · Q18

"I'd like it to tell me if I've got a chance. I know it can't really know that. But something more helpful than a higher score and then another rejection."

— P19 · Business management · Q18

"Better technical practice, probably. Show me a problem and let me work through it. I'd want to see why it marked an answer wrong. I don't need it writing another CV."

— P18 · Cybersecurity · Q18

"I'd like to speak to someone doing one of these assistant jobs. Just ask what their day is like. The descriptions don't help me picture it very well."

— P13 · Business and entrepreneurship · Q18

Another group asks for something different entirely — access, timing, and support that fits a life.

"Evening appointments would help. They don't all have to be face to face. I could talk through a form after I've finished everything else. I don't need a new app for that, I don't think."

— P15 · Business management · Q18

"Audio, and simpler text. I'd like to practise with the video timer too. That's the bit the chat doesn't help with."

— P09 · Business and computing · Q18

"It would have to take my shifts and my daughter into account. And the care qualifications. I wouldn't want to keep explaining all that every time."

— P07 · Health and care management, online · Q18

"I'd want to put in the job and find out what I need for it. I'm looking at prisons and probation and it's not all the same. And when it changes my English, explain why."

— P08 · Criminology and law, online · Q18

These are not versions of the same request. A design answering only the first group leaves the second where it already is.