In the spring of 2023, the most common AI-assisted interview prep was essentially: paste the job description into ChatGPT and ask it to generate sample questions. People treated the output as a cheat sheet and memorized answers off it. This worked, sort of, in the way that cramming flash cards works, right up until the interviewer went off-script.
The category has moved since then. The question of whether AI is genuinely changing how candidates prepare, or whether it is mostly adding new surface area to the same old anxiety, is worth thinking about clearly.
What the pre-AI baseline actually was
Before AI tools, the standard advice was: practice STAR stories, research the company, do a mock interview with a friend. Most people did roughly none of this with any seriousness. The “mock interview with a friend” advice in particular has always been difficult in practice. Finding someone willing to spend an hour on this, who knows enough about the role to ask useful follow-ups, and who will give honest feedback without making things awkward, is genuinely hard.
The result was that most interview preparation was solitary and passive. Reading about the company. Mentally rehearsing answers without saying them out loud. Hoping for the best.
There is a real problem in this, and it is not that candidates were lazy. It is that the available practice tools were genuinely limited. Recorded mock interviews were expensive. Peer practice was logistically difficult. The gap between knowing what you should say and being able to say it fluently under pressure was hard to close without repetitions.
Where AI prep tools actually help
The clearest benefit is accessible repetitions. You can run a mock interview at 11pm on a Tuesday without scheduling anyone. You can run it three times in a row if the first two felt rough. This is genuinely different from what was available two years ago, and I think it matters for a meaningful number of candidates.
The second benefit is feedback specificity. A good AI mock interview tool does not just say “your answer could be stronger.” It can tell you that your STAR story had a vague action section, that you said “um” 14 times in a five-minute answer, or that your answer to a behavioral question was 40% situation and 10% result. That kind of granular feedback is hard to get from a human friend who is being polite.
The third benefit is breadth. AI tools can simulate interviewers from different companies, different roles, and different interview styles. A candidate going in for a Google L4 SWE role and a Series A startup engineering role should prepare somewhat differently. Tools that can model these differences are more useful than generic prep.
Tools like Craqly’s interview copilot operate in real time during practice sessions, which compresses the feedback loop compared to reviewing a recording afterward. The effect is similar to getting coaching during the practice rep, not after it.
Where the claims outpace the reality
A lot of the marketing in this category leans on statistics like “over 90% of Fortune 500 companies use automated screening.” That may be technically accurate but it is doing a lot of work. Using any automated tool in the hiring pipeline, including basic ATS keyword filtering, gets counted in that figure. The actual experience of candidates varies enormously by company, role level, and how seriously the recruiting team takes the AI-generated screening results.
The Unilever case gets cited often: they reduced their hiring process from four months to two weeks using AI video screening. This is a real case. It is also a very specific context, high-volume entry-level hiring at a global consumer goods company, that does not map onto most job searches. Whether or not AI video screening is coming for your specific interview process depends on the company’s recruiting scale and philosophy, not a general trend.
I am also skeptical of the claim that AI prep tools close the confidence gap for candidates who have interview anxiety. Repetitions help. But anxiety in a real interview comes from the stakes, the uncertainty, and the social evaluation dynamic, and those are not present in a simulated session the same way. Candidates who feel very comfortable in their AI prep sessions have sometimes been surprised by how differently the live interview feels. This is not a knock on the tools; it is just something to account for.
The companies-using-AI angle is real but overhyped
Hiring teams at many large companies are now using AI to screen resumes, score video interviews, and in some cases flag candidate responses for review. This is true. The implication some career advisors draw, that candidates should therefore “prepare for AI screening,” is mostly noise.
The actual practical advice is the same as it was before AI screening existed: answer what was asked, be specific, use clear structure. The difference is that AI screeners are more literal about this than human screeners. They are not picking up on warmth or enthusiasm. They are flagging whether you addressed the question, whether your answer had discernible structure, and whether you used relevant vocabulary.
The Bureau of Labor Statistics projects that employment in most professional occupations will continue growing through 2032 (BLS Occupational Outlook Handbook). More jobs means more interviews. The AI-assisted prep category is going to keep growing simply because there is demand, not because the tools are dramatically better than alternatives.
What probably has not changed, and should not
The fundamentals of a good interview performance have not changed much. Specific stories beat general claims. Questions that show you have done real research land better than questions that show you know how to use Google. Genuine interest in the role comes through and its absence comes through.
AI prep tools add a practice layer that is genuinely useful. They do not substitute for understanding the company, knowing what you actually want, or being honest in the room about what you have and have not done.
If you asked me in 2023 whether these tools would be meaningfully changing outcomes, I would have said probably not. I have updated on that. The access to quality repetitions at low cost is real. What it changes is the floor of preparation quality, not the ceiling.
The candidates who are doing the best in this environment are using AI prep tools to do more quality reps, not to find shortcuts. That distinction matters more than which specific tool they are using.