That rule turns out to be the most useful finding in the whole exercise.
The category splits in two, and the split is about you
Every vendor here does roughly similar-sounding things. The line that actually predicts whether you can buy the product is whether they publish a price.
Group 1 — quote-only (enterprise sales motion)
Verified 20 July 2026:
- HireVue — names Essential and Premium packages, publishes no prices (source)
- Greenhouse — Core / Plus / Pro; pricing "customized based on your company's hiring needs" (source)
- Sapia.ai — per-hire pricing; states it is "designed to best serve companies that hire more than 500 people per year" (source)
- Ashby — publishes its entry tier, but Plus (101–1,000 employees) and Enterprise (1,000+) are quote-only (source)
You will find dollar figures for these vendors circulating online. We have deliberately not repeated any of them: every figure we could trace led back to a competitor's blog rather than to the vendor. Publishing an unverified price about a rival is unfair to them and worthless to you.
Group 2 — published prices (self-serve motion)
| Vendor | Published price | Volume included |
|---|---|---|
| Manatal | $15/user/mo annual · $19 monthly (Professional) | Up to 15 jobs, 10,000 candidates |
| Metaview | Free · $100 · $300 per user/mo | 100 profiles free / 200 per month / unlimited |
| Willo | $279/mo (Growth, monthly) · $209/mo annual | 2 assessments/month, up to 4 users |
| Workable | $299/mo Standard · $599 Premier · $719 Enterprise | Unlimited active jobs |
| Ashby | $400/mo (Foundations) | Up to 100 employees |
| NiceHire | $99/mo Team · $299 Business · $999 Enterprise | Unlimited users; 60 / 200 / 600 AI-interview minutes, metered above |
Sources: Manatal · Metaview · Willo · Workable · Ashby
Why this matters more in Hong Kong than almost anywhere
Hong Kong's employer base is overwhelmingly small. Per the Hong Kong Trade and Industry Department: as at March 2026 there were around 360,000 SMEs in Hong Kong, accounting for over 98% of the total number of enterprises and providing job opportunities to about 1.2 million persons — more than 43% of the total number of persons engaged. The official definition treats non-manufacturing enterprises with fewer than 50 persons engaged as SMEs, and manufacturing enterprises with fewer than 100. (source)
Read that against the two groups above. Essentially every Hong Kong employer is an SMB buyer. An enterprise sales motion calibrated to companies hiring 500+ people a year is not under-serving this market by accident — it structurally cannot reach it. The economics of a six-month sales cycle do not work against a company hiring twelve people.
That is the gap this page exists to describe.
But we should not pretend the published-price group reaches all of it. "SME" in the official definition runs from two people to forty-nine, and price segments that band hard. At the bottom of it, the realistic options are an inexpensive ATS (Manatal at $15/user) or no tooling at all — a five-person firm hiring three people a year should not be buying an AI first round from anyone, including us.
NiceHire's entry plan is $99/month with 60 AI-interview minutes included, which is the lowest flat price in the table above and does reach further down that band than the rest of the group. It is still a real budget line, and it is still aimed at teams with enough hiring volume that the first round has become a bottleneck.
So the honest version of the claim is narrower than "the quote-only vendors ignore HK SMEs, and we don't." It is: the quote-only vendors cannot reach any of this market, and the published-price vendors reach different slices of it. Ours is the slice where volume, not budget, is the binding constraint.
These products are not doing the same job
Price comparison is misleading until you separate what the tools actually produce.
Applicant tracking systems — Manatal, Workable, Greenhouse, Ashby. They store, organise and track applicants. Increasingly they bolt AI features on top. But the core job is record-keeping and workflow: the judging is still done by a human. Workable, for example, publishes video interviews as a $109/month add-on to its $299 Standard plan, and meters AI credits separately (5,000 credits for $600).
AI notetakers and sourcing assistants — Metaview. Worth being precise, because Metaview is often mislabelled in comparison tables: it is not an AI interviewer. It records and summarises interviews that humans conduct, sources candidates, and reviews applications. If your problem is that your interviewers take bad notes, that is a genuinely good product and this page is not an argument against it.
One-way video assessment — Willo. Candidates record answers to preset questions; your team watches them back. Real time savings on scheduling. Note the volume shape: the $279/month Growth tier includes two assessments per month. Scale, at $409/month monthly ($307 annual), includes ten. For a small number of senior hires that is fine. For screening sixty graduate applicants, the maths does not work.
Enterprise AI interviewing — HireVue, Sapia.ai. Deep, mature, operationally serious platforms built for very high volume. Sapia states its own threshold: 500+ hires per year.
AI first round with a ranked output — NiceHire. Resume vetted and scored against the role; a structured first-round interview drawn from the same question set, in the same order, for every applicant to that role; a transcript and a structured evaluation per candidate; a ranked shortlist. Each stage has a passing threshold, and auto-reject below it is off by default, so a person reviews every below-threshold candidate instead; it can be switched on per stage.
Where each one is the right answer
We would rather you buy the correct thing than buy ours.
- Budget is your binding constraint. Manatal at $15/user/mo is the cheapest thing on this page by a wide margin, and under about seven seats it is cheaper than NiceHire's $99 entry plan. We are not going to argue you out of that.
- You have no system of record. Buy an ATS first — Manatal is inexpensive and APAC-native; Workable is more complete and more expensive. An AI first round on top of chaos is the wrong order.
- You hire 500+ people a year across multiple markets. Talk to HireVue or Sapia.ai. That is genuinely their competence and they have been doing it far longer than we have.
- Your interviews are fine, your notes are not. Metaview, and nothing else on this list.
- You do a handful of senior hires and want to skip scheduling. Willo's one-way format is a clean fit at that volume.
- You are 101–1,000 employees and want one integrated recruiting system. Ashby's Plus tier is worth a quote.
- Your first round has stopped scaling and you cannot explain your screening decisions. That is what NiceHire is for.
The question the category answers badly
Every vendor here sells speed. Speed is table stakes, and for most buyers it is also the stated problem rather than the real one.
The real one surfaces later, and it sounds like this: a rejected candidate, your CHRO, or a regulator asks why someone did not progress — and nobody can produce an answer. Screening that happened in a recruiter's head across 300 CVs on a deadline leaves no record. It may well have been good judgment. It is simply not evidenced.
A structured first round leaves a record by construction: the same question set for every applicant to a role, a stored transcript of the interview, and a structured score. What that record does and does not support is set out below.
A note on compliance, stated carefully
Regulatory attention on AI in hiring is rising, and vendors in this category are not always careful about what they promise. So, precisely:
NiceHire cannot make you compliant with any bias-audit law, and no vendor can. Where an independent bias audit is required — New York City Local Law 144 is the clearest example — the audit must be performed by an independent auditor. A vendor's own assertion is not a substitute and never has been.
What a structured first round leaves behind is a record, and we would rather describe ours accurately than call it something grander. A completed screening stores the interview transcript against the application, along with the scores, the recommendation, and the recorded strengths and gaps. Every applicant to a role is asked from the same question set, in the same order.
Three limits on that record, because they matter to anyone actually planning to rely on it. The transcript is whatever the interview provider returns, so a session that ends early can be stored without one. Re-running a screening clears the previous attempt's transcript rather than keeping both. And the interview template is stored as a live reference rather than a snapshot, so editing a template changes what the record says about interviews that already happened.
Those are real gaps and they are on our list. We are not going to describe this as an audit trail until they are closed.
What it is good for today is the ordinary version of the question: when someone asks why a candidate did not progress, there is something to look at other than a recollection.
On being local
NiceHire is built in Hong Kong by VecTech Limited. The interface ships in English, Japanese and Simplified Chinese; Japanese coverage is the least complete of the three and falls back to English rather than showing an untranslated key. We ingest live job postings from Hong Kong and Japan employers, including a large number of financial institutions, which is where our reading of this market's hiring patterns comes from.
That is a genuine advantage over a vendor treating APAC as a later localisation project, and we will not overstate it beyond what is true today. The long-form comparison pages on this site, including this one, are currently written in English only.