AI interview services: 14 compared (October 2026)

Conclusion first

Key takeaways

Start here

What an AI interview is

An AI interview is one in which the candidate talks with an AI. They open a link and take it without waiting for an interviewer to be free. The AI asks questions set in advance and follows up on the answers. It then records a score against each criterion, with the reasoning behind it. People decide who is hired.

Taken together, comparison articles cover more than twenty services available in Japan. Most offer much the same features, so the more services there are, the more work it takes to compare them.

Background

Why AI interviews are spreading

Vendors tend to name the same two problems.

Many AI interview services say these are the two problems they set out to solve.

Candidates feel it too: while they wait for an interviewer to become available, another company’s process moves ahead.

The solution is also much the same everywhere: make the interview available 24 hours a day, 365 days a year and run any number of them in parallel. Nothing has to wait for an interviewer to be free, so both the workload and the waiting time fall. Up to this point, every service makes the same pitch.

What changes

Benefits of AI interviews

The more applicants you have, the greater the benefit.

Caveats

Drawbacks of AI interviews, and what to settle first

AI interviews have weak spots too. Four apply, to a greater or lesser degree, to every service.

Sources: Leverages (NALYSYS), survey of 328 students graduating in 2027 to 2029, May 2026 (Mynavi News, in Japanese); Preferred Networks, survey of 308 students graduating in 2027 (30 September 2026, in Japanese). Both companies provide AI interview services.

The first two can be reduced, to a point, by revising the questions based on the interview records. With Proomi, Prance can also adjust finer points case by case, such as how deeply the AI probes and how it paces the conversation (at additional cost, quoted separately). See “Question design and AI interviewer behaviour”.

With that in mind, settle the following before you start.

The problem

Why comparisons rarely settle the choice

The feature grids fill with the same rows, most vendors keep pricing off the site, and “you can tailor the evaluation criteria” appears on every one. There is more to compare, but less to base a decision on.

One reason is the axis used for comparison. Most comparison articles sort AI interview services by how much the AI takes over: conversational, one-way recorded, or support for recording and analysing interviews. That axis tells you what the AI does. It does not tell you who owns the evaluation criteria. Who owns them determines how much you can adjust and what you can do a year after adoption.

Classification

How AI interview services differ

So we sorted them by what each service optimises for. This gives the four types below.

The evaluation items to put in a comparison table are set out in how to choose an AI interview tool. Here we focus on the basis for assigning each type. Whether the criteria can be written freely is a separate matter from the type: a service that lets you write criteria freely but leads on low price, for example, sits in the low-cost, flat-rate type. Within each type, we then look at three things: the questions, the criteria and the verification.

We do not rank them. Which type is right depends on what you want to optimise. Proomi is simply one of the four.

The table

Where 14 services fall

Source: public information from each vendor, checked on 22 October 2026. The pages we used are listed at the end.

Choosing between them

What each one optimises for

01Shared-framework type

The vendor supplies the framework that evaluation is built on. PeopleX builds on around 150 preset questions and lets you add your own; it also states that you can “design evaluation items freely”, and the “weighting” of each item can be changed. SHaiN (Talent and Assessment) is built on its “strategic hiring method”: evaluation draws on 18 supplied items, while up to 20 free questions can be added. TA Team (VARIETAS) also supplies its 20 observation points and reweights them against the role profile; a client release states that VARIETAS “develops the interview design to match the items the company wants to evaluate”.

TG-WEB mee (Humanage) scores competencies from past behaviour, using an “evaluation logic based on aptitude-test data from five million people”. The service supplies the evaluation items, and a company adds its own standards through free questions. NALYSYS AI Interview (Leverages) has you either pick an evaluation model for new-graduate, mid-career or part-time and staffing hiring, or build the interview from the items you value among 33 it supplies; either way, you can add questions of your own. Calaris Assess (Calaris) is built on its own framework, “Caliber”: it brings in the points each company wants to emphasise and, it says, adjusts the weighting of the criteria to each HR decision.

The scope of customisation needs to be checked separately for questions and for criteria. Of the six services in this type, three mention verification in their public information: PeopleX, SHaiN and TA Team. SHaiN displays the “deviation and frequency” of each evaluation item. PeopleX says it will “propose evaluation criteria” and reflect them in interview questions and evaluation items; its feature list notes that this “includes features in development”. TA Team runs as five AI agents, of which an audit agent “monitors the integrity of hiring continuously”, and it describes its approach as a method that proposes “what to change in the next round”.

02Low-cost / flat-rate type

MiAI (Props) charges from ¥1,000 per interview, a level at which you need not hesitate over volume. Our AI Mensetsu (JetB) is a flat-rate service that states “monthly fee only, no usage charges”, so the bill does not move with the number of interviews. AI RECOMEN (i-enter) states “no set-up fee and no fixed monthly fee”: you buy interviews as tickets at ¥700 each. Up to five evaluation standards can be set.

03Sourcing / pool type

CaseMatch also offers a scouting product and uses interview-practice history to build the candidate pool. Z AI Interview (ROXX, formerly Z Career AI Interviewer) began in October 2024 as a feature for companies using Z Career, a job platform for non-desk workers, and opened to standalone use in January 2025. Its release states it can be used “not only with Z Career but also for interviews arising from recommendations through recruitment agencies”. It starts from interviewing candidates the platform has gathered.

04Criteria-owned type

Here the evaluation criteria become yours. With Track, harutaka and Proomi alike, the criteria are built from your own requirements. What differs is who designs them and who corrects them.

Proomi (our own service) has the customer define the criteria, and the check on whether those criteria were right also runs on the customer’s data. The score distribution per criterion shows where the AI is separating candidates, matching against hire/no-hire results shows which criteria actually drove the decision, and following retention afterwards shows which ones held up. When the AI and the interviewer disagree, you change the definition yourself rather than waiting for a model update. Evaluation uses only what the candidate says, not facial expression, tone of voice, or speaking pace. Because the evidence sits in what was said, you can trace why a criterion scored as it did on your own data.

Track AI Interview (Givery) “designs the evaluation items, scoring criteria and question flow individually” and states that it “improves them periodically in operation, optimised to your criteria”. The criteria are described as yours, while the design work and the corrections after go-live sit with Givery.

harutaka AI Interview (ZENKIGEN) also states that “evaluation axes, question content and follow-up scenarios can be designed flexibly for each company”, with a dedicated representative supporting the interview design from adoption through operation, to match your hiring criteria. As with Track, the design work sits with the vendor.

When you look at this type, ask who does the work of writing the criteria and who does the work of correcting them once you are live. A service that assumes you will define and improve everything in-house and one where the vendor stays through design and operation mean very different workloads, even when both say “customisable”.

Even in high-volume new-graduate hiring, if you define the criteria yourself, the competency model and behavioural definitions you already have can be used in the AI interview unchanged. Scoring can be written as a rubric, so the yardstick is the one your human interviewers use.

The criteria used in selection then carry over to evaluation after joining. Use the same format with current employees and you can see from real data which criteria the strong performers score on. Compare that with retention and you learn whether that year’s criteria were right, and next year’s can be rewritten in-house without waiting for the vendor to update a model. The larger the intake, the faster that cycle runs.

By situation

Which type suits which kind of hiring

The right type is not always one choice for the whole company. It can differ from one kind of hiring to another: a company might borrow an established framework for new graduates and write its own criteria for each mid-career role. For each one below, we note what to check and, where one or two types clearly stand out, which to look at first. Each card links to a page on how Proomi is used in that situation.

What decides it

What to ask in the meeting

“Can we adapt the evaluation criteria to our company?” no longer separates the types. All 14 services claim customisation of some kind, and even a low-cost service priced from ¥1,000 per interview says you can design everything yourself. The difference sits one level finer: whether you define the evaluation items themselves or reweight the ones supplied, and how far back you can check whether the evaluation was right.

Verification is not a yes-or-no question either. Seven of the 14 say something about it, and their checks go to different depths: the distribution across candidates, the design of the next round, retention after joining. Only one says its checks extend to retention. Ask the three questions below and you will see how far each service lets you customise, and how far its checks reach.

Whichever of the three you ask, when the answer is yes, ask one more thing. Is that feature running today? Some services note that their feature list includes items in development, and public information does not tell you what has shipped.

These three questions and others are set out as items for a comparison table, with guidance on reading the answers, in how to choose an AI interview tool.

Pricing

Compare pricing by how it is structured

Many AI interview services do not publish their prices, so comparing them on price alone is difficult. If you first work out how each service sets its price, quotes are easier to compare when they arrive. These are the common structures.

In terms of the four types, the low-cost / flat-rate type optimises for the cost per interview itself. For the shared-framework and criteria-owned types, the price tends to depend on how much evaluation design and revision after go-live is included, and is often given only as a quote. To compare, work out the annual total at the same interview volume, then compare who is responsible for revising the criteria.

Proomi (our own service) is priced on an annual allowance plus overage. Starter is ¥95,000 a month and includes 240 interviews a year; Growth is ¥160,000 a month and includes 500. If you would rather start small, Flex lets you prepay for interviews at ¥4,800 each, starting from 100.

The setup fee depends on the scope of implementation support, so we provide it in a formal quotation.

On the pricing page, you can enter the number of interviews you expect each year and estimate the cost of each plan.

Rollout

How an AI interview rollout usually runs

Whatever the type or the service, the steps of a rollout are much the same. A typical order is below. If you raise the likely sticking points for each step while you are still choosing a service, there is less to redo after go-live.

There is no standard timeline for a rollout. A company whose criteria are already documented can get through “Put your criteria into words” quickly. The more candidates you put through both interviews in parallel, the sooner you have enough material for “Compare the results and revise”.

If you rewrite the criteria in Proomi, you can check before go-live how the scores of past interviews would change. The steps are in “Rewrite your criteria, then test them on past interviews”.

FAQ

Frequently asked questions

What is an AI interview?

An AI interview is one in which the candidate talks with an AI. They open a link and take it without waiting for an interviewer to be free. The AI asks questions set in advance and follows up on the answers. It then records a score against each criterion, with the reasoning behind it.

Does the AI decide who is hired?

People do. Across vendors the stated arrangement is the same: the AI goes as far as the score and the reasoning. How much is handed to it varies, so it is worth asking when you compare services. How to check the scoring is covered in “Choose an AI interview tool that can explain its scores”.

If it says “customisable”, is it the criteria-owned type?

No. All 14 claim customisation of some kind. What differs is who decides. For questions: whether the service designs them from your own criteria, or you write them yourself as free text. For criteria: whether you pick from supplied items, reweight supplied items, have the service design them from your own criteria, or define them yourself as free text. The table uses those words. Beyond that, what separates them is whether you can check that the criteria were right. Items for a comparison table, with guidance on reading the answers, are in “How to choose an AI interview tool”.

Are there services that sit across two types?

Yes. Some services offer both introductions from a candidate pool and an AI interview designed on the customer’s criteria. We classify each service by what the vendor emphasises publicly, and where a service sits across two types the table names both.

Can current employees sit the same evaluation as candidates?

Yes. No special feature is needed: if you can define the criteria yourself, you simply put the same questions and the same criteria to current employees as to candidates. Only the person in the interview changes. Where the vendor designs the evaluation items, you would ask it to design a version for employees. When employees go through the same format, you can see from real data which criteria the people who do well actually score on. This is where differences in the depth of verification show. The thinking behind it is in “Was your interview evaluation right? You find out after the hire”.

Can the same criteria run across roles and languages?

Any type can apply one standard across every role and language. What differs is whose standard it is. Where the vendor designs the framework, that framework becomes the common standard. It holds steady as roles and languages are added, but it was built separately from your own grades and competencies.

Where you can define the criteria as free text, you can apply the single standard you already have to every role and language. Even if the emphasis shifts by role, the underlying definitions stay in your own words. Because they use the same terms as your performance reviews and grade definitions, hiring and life after joining are described in the same language.

What if we cannot write the criteria ourselves?

You do not need finished criteria to choose the criteria-owned type. We can build them with you, from the definitions your performance reviews already use and from what your interviewers actually look at. What matters is that the finished criteria belong to you and that you can rewrite them afterwards. Who writes the first draft is not what separates the types. How to start from your performance review criteria is in “Use your performance review criteria in interviews, word for word”. To talk it through using your current criteria, contact us.

Is there a free trial?

With Proomi, every plan includes a free trial, and you can try a real AI interview soon after getting in touch. Whichever service you trial, check whether you can test it with your own criteria and questions; if you can, what you see in the trial should be closer to what you see after go-live.

How is pricing set?

Depending on the service, you pay per interview, pay a flat monthly fee, or pay on an annual allowance plus overage, and setup fees are handled differently too. Working out the annual total at the same interview volume lets you compare services with different structures side by side. For the differences between the structures and for Proomi’s prices, see “Compare pricing by how it is structured”.

Can we give candidates feedback on their assessment?

With Proomi, feedback can be sent to the candidate after the interview. The hiring company decides whether to send it, and we work with you to adjust what it says in line with your own policy. In its default form, the feedback leaves out the pass/fail result and the scores; it points to where the candidate’s strengths came through, quoting what they said.

Can interviews be held in English?

Yes. Proomi runs interviews in Japanese and in English, and the assessment of an English interview, with the reasoning behind it, can also be reviewed in Japanese. For other languages, please ask us. How Proomi is used with international candidates is covered in “Assess international candidates against the same criteria, in Japanese or in English”.

In short

Summary

About this article

Written by Prance Holdings, which makes an AI interview service called Proomi. We include ourselves in the comparison and do not rank anyone. The table is based on what each vendor publishes; we have not verified which features have actually shipped.

Sources

Sources used for classification