Does it actually help your job hunt? We checked.
A general AI answers every topic on earth, and averages to the US to do it. Careersy is built for one thing: a tech career in Australia and New Zealand. We put both to the questions you actually ask, to see how much that focus is worth to you. Here is the working.
The local, specific answer. Every time.
| Careersy AI | about 90% |
|---|---|
| Claude (free) | 65% |
| ChatGPT (free) | 63% |
| Gemini (free) | 63% |
How useful each answer was for an ANZ job seeker, scored against the same rubric.
The gap came from two places. Local relevance: Careersy scored on it every time; the general-purpose tools scored zero, every time. None of them mentioned Australia, local pay, or the local market unless we forced it. And voice: Careersy took a position, the others returned competent, generic lists. On the facts themselves, everyone did fine. Knowing the facts was never the differentiator. Knowing this market was.
Real questions, one candidate, one rubric.
We took the questions our own users actually ask, in their own words, straight from the app. CV feedback, tailoring, LinkedIn, how an ATS really works, how to quantify achievements you cannot fully remember. Then we put the same questions to four tools: Careersy AI, and the free versions of ChatGPT, Claude and Gemini. Same questions, same candidate, same rubric. We scored every answer and counted.
The candidate was fictional and built on purpose. A Sydney software engineer moving into a Forward Deployed Engineer role, with three real gaps against the job, so we could see which tools caught them.
We ran Careersy on its normal setup and the competitors on their free tiers: ChatGPT logged out, Gemini on Flash, Claude on its standard model. That is the real choice a job seeker faces, pay for the specialist or use the free generalist. It is not a like-for-like model test, and we say so below.
Four things, scored 0 to 2. Eight points an answer.
Local relevance
Does it know the ANZ market, or average to the US?
Groundedness
Are the facts right, and does it correct the common myths?
Actionability
Concrete next steps, or vague encouragement?
Coaching voice
Does it take a position, or hedge?
The one that tells the whole story.
We handed all four the same CV and the same job. Careersy ran an actual scored read: 87 out of 100, with the three missing keywords named. ChatGPT read the same CV and said 93. It has no scoring engine. It invented a number that felt encouraging. The person who believes that 93 walks in thinking they are ready, and no one corrects them.
The holes, because you should know them.
We are not going to pretend this is a peer-reviewed study. One rater scored the answers against a fixed rubric, not an independent panel. We kept the full answer log so anyone can re-score it. It was six advice questions and three capability tests, not three hundred.
And Careersy runs a premium model while we tested the competitors on their free tiers. That is the real choice a job seeker makes, pay for the specialist or use the free generalist, but it is not a like-for-like model test. We are showing you the holes on purpose. A number you cannot check is not worth much.
A bigger question set, re-scored independently and blind. We will update this page when we do. Until then, this is the honest version: a clear, repeatable read that points one way, with its limits in plain sight.
See the full Careersy vs ChatGPT and Claude comparison.
See what a recruiter sees. Free.
Start a chat. Get a recruiter’s read on your CV, your brand, your next move. The specialist we just tested, on your actual search.