Features
Seven features,
in the order hiring happens
Write the job. Take the applications. Talk to the candidates. Decide who to meet. Hitoshiru's features are joined up along the first round rather than listed beside it.
01Job creation
The posting is where screening starts.
The role, the work, how the interview runs, what the interview is for, and the criteria with their weights, all on one screen. The introduction the candidate hears and the question suggestions are both written from what you put in the description.
- You say what the interview is for — verifying the experience, how they behave at work, why this job and will they stay, a first-round screen, communication in the role, or potential — and the suggested criteria and questions, the follow-ups and the report's summary are written for it. Up to two per job
- Criteria are scored 0–5, and a criterion's weight is how much it moves the overall
- A job cannot be published until it has criteria, weights and questions — nor while an Employment Security Act warning is unresolved
- Editing a published job starts a new version; interviews already taken keep the version they were taken under
Live: rubric v2. Editing here starts a v3 draft. Candidates get v2 until you publish.
The role
Job title
B2B sales (Tokyo, mid-career)
Job description
Selling IT services to small and mid-sized companies. Mostly inbound leads, and one person carries an account from the first visit through to the contract. Growing existing accounts is part of the role.
The introduction the interviewer gives the candidate is written from this, and so are the question suggestions.
How the interview runs
Interview language
Maximum length (min)
20
Start with a practice question
What this interview is for
Up to two. The suggested criteria and questions, the follow-ups in the interview and the report's summary are written for it.
- Verify the experience
- How they behave at work
- Why this job, and will they stay
- First-round screen
- Communication in the role
- Potential
Rubric (draft v3)
- New business2 questions ask about thisWeight330% of the overall score
- Meeting a target1 question asks about thisWeight330% of the overall score
- Coordinating inside and out1 question asks about thisWeight220% of the overall score
- Understanding the customer1 question asks about thisWeight220% of the overall score
- The posting
- AI suggests
- A person approves
- Used in the interview
4 suggested questions
AI suggestions, written from this job's description.
Nothing is added until you tick it. These are written from the job description and your scoring criteria, and each one was screened against the 職業安定法 14 topics.
Tell us about winning a new account — from the first contact through to the contract, as concretely as you can.
For New business
If there was a period when you missed your number, what did you conclude the cause was, and what did you change the following quarter?
For Meeting a target
When a customer wanted something that cut against your own company's policy, how did you settle it?
For Coordinating inside and out
When you put a proposal together, what do you find out about the customer's industry, and how far do you go?
For Understanding the customer
1 more was withheld by the 職業安定法 screen.
02AI question suggestions
What to ask, suggested.
The model writes draft criteria and interview questions from the description, the rubric and what you said the interview is for. A suggestion never goes into an interview as it stands.
- Suggestions are stored nowhere, and only what you tick is added. Nothing is ticked when the panel opens
- One question can be replaced on its own, with a fresh suggestion for that criterion, and a whole-job draft can replace what is there or add to it — your choice, and Undo before you save
- They are screened against the Employment Security Act's 14 topics; anything that matches is not shown
- Once added, a question can be edited freely — and it reaches an interview only after you publish
03Applicant management
Take the application, and carry straight on.
Each job gets its own application URL, and what arrives lands in one list. Who goes to an interview is decided by a recruiter, after reading what they sent.
- The URL goes on your own site, careers page, social or a printed advert, and anyone who opens it can apply
- An application carries the name, the contact details and their own account of themselves, plus an optional CV (PDF or Word)
- The model reads that document against the criteria and sets out what it shows, what it does not, and what to confirm. It does not score it
Application link
A link for your own careers site, a message to a list, or a printed advert. Anyone who opens it can apply to this job.
https://acme.hitoshiru.com/en/apply/7f3a…This job's application link is live.
Posted on: careers page, social, job boards
Name
Misaki Sato
Email address
m.sato@example.com
About you / your experience
Four years selling to companies, mostly working with…
CV / résumé (optional)
misaki-sato-cv.pdf
I have read the above and agree to my information being used for this application.
B2B sales · 3 applications
- Move to interview
Misaki Sato
Four years selling to companies · CV attached
- Hold
Ken Takahashi
Moving from consumer sales to selling to companies
- Not handled
Nguyen Thi Lan
Sales support at a trading company, Japanese N2
Only the people you move to interview are sent an interview link.
04The AI interview
A first round that is easy to take.
The questions a recruiter approved are put to the candidate by voice. The interviewer does not only read them out — it listens to the answer and follows up.
- No app
- Works on a phone
- Any hour, until it expires
- Interviewer
When the night shift was short-handed, what did you do?
- Candidate
I talked to a senior colleague, and the two of us changed the order we did the rounds in.
- Interviewer
After you changed the order, what changed for the residents?
A follow-up the model wrote from the answer. Screened against the 14 topics before it was spoken.
- A thin answer is followed up on the spot, and a pause for thought is not interrupted
- A follow-up is screened against the 14 topics before it is spoken, like every other question
- No camera, and no analysis of face or emotion. A dropped connection resumes where it stopped
05The candidate report
Know the candidate without watching it back.
A score and its justification per criterion, the words behind each one, and the full transcript and audio — on one page.
Three years at a special nursing home, with hands-on care and night shifts, and examples given concretely throughout. What stands out is being able to describe a mistake and the correction that followed.
Strengths
- Hands-on care and night-shift experience
- Describes a mistake and the fix in their own words
Areas to improve (what to confirm when you meet)
- Few concrete examples of records and handovers
- Only one example of working across professions
Bar length is the score, thickness is the weight. The line is the median for this role.
- Communicating with residentsStrong4.5
Weight 22% · role median 2
- Care experienceStrong4.5
Weight 33% · role median 1.5
- Safety and hygieneStrong4.0
Weight 22% · role median 1.5
- Working as a teamStrong4.0
Weight 22% · role median 1
Japanese proficiency (estimated from the conversation)
JLPT does not test speaking, so this is shown as an equivalent band.
CEFR (speaking)
JLPT equivalent
N2–N1 equivalent
Confidence
Medium
Assessed from 1 min 55 s of speech. Confidence is Medium because there was not much of it.
Quoted evidence — 12:04
「最初は止めようとしましたが、余計に興奮させてしまって。それで先輩に相談したら、一緒に歩いてみたらどうかと言われました。」
“At first I tried to stop her, but that only agitated her more. So I asked a senior colleague, who suggested walking with her instead.”
A recruiter can overwrite any score with a reason. The rubric version, the prompt-pack hash and the model version are written to that interview's audit log.
- Strengths
- What the candidate evidenced, restated in a line each
- What was thin
- Where the interview did not reach, and what to confirm when you meet
- Scores and weights
- Each criterion the job set, 0–5. Bar thickness is weight
- The evidence
- The words behind a score, quoted rather than summarised
- The evidence for a score is the candidate's own words, quoted rather than summarised
- Any score can be overwritten with a reason; the model's number and the override are kept in the audit log
- Hitoshiru never decides an outcome. The final call is a person's
06Multilingual interviewingWhen you need it
A different language. The same person, in full.
The interview language is set per job — Japanese, English, or both, in which case the candidate chooses.
- An interview conducted in English still arrives as a Japanese report
- The quoted evidence keeps the original words alongside the Japanese
- One interview is pinned to one language, and the interviewer's voice is the same throughout
Interview language
日本語
This job's default
English
The candidate may choose it
Where a job offers both, the candidate picks at the start of the interview. One interview is pinned to one language, and the interviewer's voice is the same from beginning to end.
The report comes in Japanese
English → Japanese reportAn interview conducted in English still returns its scores and justifications in Japanese. Only the quoted evidence keeps the original words.
“At first I tried to stop her, but that only agitated her more. So I asked a senior colleague, who suggested walking with her instead.”
「最初は止めようとしましたが、余計に興奮させてしまって。それで先輩に相談したら、一緒に歩いてみたらどうかと言われました。」
An illustration. Not a real candidate.
JLPT equivalent band
N2–N1
equivalent, read from the conversation
CEFR (speaking)
B2
Confidence
Medium
JLPT does not test speaking, so this is shown as an equivalent band. Assessed from 1 min 55 s of speech; confidence is Medium because there was not much of it.
- Fluency
- Few hesitations, and after getting stuck they recover and carry on unaided.
- Accuracy
- Errors in particles and tense, none of which gets in the way of being understood.
- Range
- Concrete vocabulary for care work. The topics stayed within the job.
- Interaction
- Handles an unexpected question by asking back, and turn-taking is natural.
- Phonological control
- Intelligible throughout. Accent is not assessed.
Quoted evidence — 08:41
「夜勤のときは、申し送りで気になったことを必ず自分の言葉で言い直すようにしていました。そうしないと、次の人に伝わらないので。」
07Japanese proficiencyWhen you need it
"They can speak Japanese", made more specific.
Read from the conversation of an interview conducted in Japanese. There is no separate test, and the result is never a single JLPT level.
- A band such as "N2–N1 equivalent", with a confidence level and the speech it rests on
- Fluency, accuracy, range, interaction and phonological control, described in sentences rather than scored
- An interview conducted in English contains no Japanese speech, so nothing is assessed
One flow
Seven features. One flow to use them in.
- Job creation
- Question suggestions
- Applicants
- AI interview
- Candidate report
- A person decides
Five features join up the first round. Two join in only where the hiring calls for them.
When you need it
- Multilingual interviewing
- Japanese proficiency
Not stages every candidate goes through. The interview language is set per job, and an interview conducted in Japanese can have the candidate's Japanese read from the conversation.
Alongside
Not only features — something you can run hiring on.
Multi-tenant operation for staffing agencies
Many client companies, their jobs and their candidates in one console, with each member's permissions alongside. Looking across companies is blocked by the design, not by a policy.
Read more →The screening record and the audit log
Who was asked what, how they answered, and why the score is what it is — down to the questions that were discarded — in an append-only, hash-chained log. None of it can be rewritten afterwards.
Read more →How candidate data is handled
Recordings stay on our own servers, encrypted under a separate key each. Interview links are stored as hashes only, and a deletion request is answered in a form nobody can undo.
Read more →
See it for yourself
Bring a job posting and we will draft its criteria and questions on the call. About thirty minutes covers both the candidate's screen and the report.

