Can the reviewer tell AI did my task submission?
Yes, most of the time they can. It is rarely the polish that gives it away. Task submissions are read by the technical team who own the problem you would be hired to solve, and a submission that AI carried alone tends to be complete on structure and thin on judgement. If it still gets you to the next meeting, expect to be asked why you chose each part of it.
So the real question is not whether you used AI. It is whether anyone can still find you in the work.
Who actually reads my task submission?
The technical team reads it, and they read it properly. In most processes I have seen, the task is built around something relevant to the organisation, so the reviewers are not marking an exam. They are asking whether your approach could work for them. HR is sometimes involved and adds its own scoring, but the decision sits with the technical side.
Because the task mirrors a real problem, shallow means something specific to that team. A shallow submission does not read as a lazy candidate. It reads as a candidate with not much to add to the problem they are hiring for. That is a much harder conclusion to come back from.
Why does an AI-built submission feel so convincing to the person who sends it?
I used AI on my own task submissions too, to speed things up, and at first I was a little fooled by it. It is very good at showing you angles you did not think about, and the structure it gives you makes you feel reassured, almost reliant on it. Then you go deeper into the content, and the gaps start to show.
One of mine was a hiring strategy task. On paper the AI plan was great: channels, stages, timelines, all well organised. What it forgot were the things that often work better than any of that: the company's own employer brand, referrals and word of mouth. AI was strong at the structure and weak at the human angle, the part that comes from having done the job and seen what actually moves people.
A senior reviewer notices exactly that kind of gap, because they would never have left it themselves. And like with AI slop on your feed, once they notice, they quietly detach and move on.
What happens if a shallow submission still gets me to the meeting?
It happens, though rarely. Some reviewers invite the candidate precisely to test how deep the knowledge behind the submission goes, or to give them a chance to clarify. Either way, your submission becomes the agenda for that meeting.
I have seen many developers fail at this point. The question was simple: why did you use this approach, or why this code? And the answer was not there, most likely because the AI had made that decision, not them. It is the same weakness that sinks confident answers in interviews: the claim holds until somebody asks the second question.
Think of it as karaoke. The backing track can be perfect, but the room is waiting to hear whether you can actually sing.
I call this the follow-up test, and it is the idea at the heart of Not in the Room: the candidates who pass are the ones whose evidence still stands after the second and third question.
How much effort does a task submission really need?
More than the brief says. Most tasks I took on estimated 2 to 4 hours, and I invested close to double, and that was before AI was as advanced as it is today. All of them got me to the next stage. I would not credit the hours alone, but where they went. AI gave me a strong foundation, then I spent my time on what it missed, on the language, and on tailoring the presentation to the company's own tone and design so it read as mine.
Is it worth it? When the volume of applicants keeps growing and everyone fights over the same spot, only you can answer that. The worst case is a few hours lost. In my experience, every task I took on, even the ones that took days, left me more capable than before.
Should I tell them I used AI?
Yes, if your expertise is genuine. Most organisations do not restrict AI, but they do assume that whatever you hand over represents you: the knowledge, the language and the way it is presented. A short note on how you used AI, what it helped with and where your own judgement took over, shows ability to think, trustworthiness and transparency. Those are the qualities hiring managers are looking for in the first place.
I have not yet seen a hiring manager react badly to a note like that. It is still not common practice, which is exactly why it makes you stand out.
AI can deliver your message. It cannot do your thinking, and your thinking is the only part the reviewer is hiring.
Nobody puts the actual reason in the rejection email.
Not in the Room shows who really decides, why strong candidates lose on trust rather than skill, and what a rejection email actually means. Worth reading before the next application goes out. €5.
Comments