What an AI screening report looks like
When Blue finishes interviewing an applicant, this is what the hiring team reads: a short summary, the answers it was drawn from, one score for the whole conversation, and the transcript underneath. Below is a complete example for a retail assistant opening, with a shorter one for a sales role.
Illustrative example. Every name, answer, score and figure on this page is invented. Agentic Blue publishes no candidate data.
Describe your own role and Blue drafts the questions it would ask.
The screening summary
Blue writes the summary from the interview, in a few lines a recruiter can read between calls.
- Fourteen months on the floor at a mall clothing outlet, including closing shifts and the weekly stock count, which the supervisor let them run alone.
- Wants RM 2,000 a month and says it is negotiable if the shift is fixed. One week of notice, currently part-time.
- Can work weekends and public holidays, and cannot start before 11am on Mondays.
The answers it was drawn from
The four things a first round has to establish, kept as the candidate gave them. An answer that never came up is marked “Not covered” rather than guessed.
- Relevant experience
- Fourteen months at a mall outlet, closing shifts and stock counts
- Expected salary
- RM 2,000 a month, negotiable for a fixed shift
- Notice period
- One week, currently part-time
- Availability
- Weekends and public holidays, cannot work before 11am on Mondays
The score, and what it is not
82 out of 100, for the whole conversation. There is no pass mark and no per-question verdict: Blue returns one score against the criteria you set for the role, and the reasoning sits beside it as strengths and concerns.
Strengths
- Closing and stock-count experience, unprompted and specific
- Pay expectation inside the range for the opening
- One week of notice
Concerns
- Cannot start before 11am on Mondays, which the morning shift needs
- No experience on a point-of-sale system of the kind this outlet uses
Answers are recorded as given, not judged. The score reflects the whole conversation, so the transcript is part of the read.
The transcript, including the follow-up
Every word is kept. This excerpt ends where it matters: a vague answer about pay and start date earns a follow-up, and the summary above could only be written because of it.
Blue
Before we start: I am an AI interviewer, and a person at the company reads your answers afterwards. Ready when you are. What have you done that is closest to retail work?
Candidate
I worked at a clothing shop in a mall for over a year. Mostly the floor, sometimes the counter.
Blue
What were you responsible for at closing?
Candidate
Cash up, tidy the racks, and the stock count on Sundays. My supervisor let me do the count alone after a few months.
Blue
What pay are you looking for, and when could you start?
Candidate
Depends on the shift lah. Can start quite soon.
Blue
That is fair. To put a number on it for the hiring team: what monthly figure would you accept for a fixed shift, and how much notice do you owe your current employer?
Candidate
RM 2,000 if the shift is fixed. I am part-time now so one week notice is enough.
The same role, ranked
Finished interviews are ordered against the criteria you set, with the reasoning attached. Blue rejects nobody: the order is a reading order, and the hiring decision stays with your team.
- 82Nurul HakimClosing experience, one week notice, pay in range
- 71Lim Wei HongRetail experience, three weeks notice, pay above range
- 66Daniel FooNo retail experience, available immediately
Illustrative example. Every name, answer, score and figure on this page is invented. Agentic Blue publishes no candidate data.
A shorter one: sales executive
The parts do not change with the role. For a sales opening the same report reads: six years in B2B software sales, four of them closing mid-market accounts in Malaysia and Singapore; RM 11,000 a month plus allowance; one month of notice; available for a call any weekday after 6pm. Score 79, with a concern that the largest deal named was carried by a team of five rather than alone.
What the recruiter does next
Read the top of the list, open the transcripts for the ones worth a conversation, and arrange your own interview with those. The questions worth asking in that second round are in the first-round questions guide, and the screening questions for each role are on the role pages.
What is not in the report
- No score for a face, a voice or an accent. Blue reads the words.
- No identity check and no screen watching. We do not claim to detect a candidate who had help writing an answer.
- No per-question pass or fail, and no automatic rejection.
- No personal identifiers pulled from the CV: IC and passport numbers, bank details, date of birth and address are removed before Blue reads it.
How the data is handled is on the trust page, and prices are on the pricing page.