10 Questions to Ask in an ATS Demo (APAC Edition)

A vendor demo is designed to go well. The script runs on clean data, the presenter drives, and every feature works on the first click.

Your job in the room is to break that script politely: ask the questions that make the vendor prove a claim on your terms, and listen for the answer that reveals whether the platform fits APAC hiring or only looks like it does.

These ten questions come straight from the things that decide whether an ATS works across Southeast Asia and Hong Kong. For each one, this guide gives you why it matters, what a good answer sounds like, and the red-flag answer that should lower the score. It is written for the person running the demo with a shortlist already in hand.

Before the demo: what to bring

Vendors demo on their data. You learn more when they run on yours. Bring:

• Real resumes from three of your markets, including at least one in a local language
• One live job opening you are actually hiring for
• Your markets' basic data requirements, so the compliance answers are concrete

Tell the vendor in advance you will want to see the AI run on your files. A vendor that resists is already answering a question.

The ten questions

1. Can you parse real resumes from three of our markets, one in a local language?

Why: AI screening quality collapses on local-language and messy resumes, which is most of what you receive.

Good answer: accurate field extraction with no manual cleanup, on your files, live.

Red flag: "We can set that up after onboarding," or a parse that only works on the clean English sample they brought.

2. What happens in WhatsApp after the 24-hour window closes?

Why: outside the 24-hour customer service window, you can only reach a candidate with an approved template, and WhatsApp is the frontline channel across the region.

Good answer: approved templates, plus an automatic email fallback.

Red flag: no awareness of the window, or no fallback when it closes.

3. Where is candidate data stored, and can retention be set per country?

Why: each APAC market sets its own rules on candidate data and cross-border transfer, and this is a go/no-go for legal.

Good answer: a named hosting region and retention that can be set per market.

Red flag: a vague "in the cloud," or one global retention setting for all markets.

4. How do you calculate the AI-to-recruiter agreement rate?

Why: the headline accuracy number is only meaningful if you know how it was measured.

Good answer: a stated sample size, roles, markets, and method.

Red flag: a round number with no method behind it, or "customers see around 90%."

5. Who makes the final hiring decision?

Why: in several markets, how AI is used in hiring is a live regulatory question, and a person should own the decision.

Good answer: a person decides, with the AI's reasoning logged.

Red flag: the AI described as the decision-maker, or no record of its reasoning.

6. Can you run a first-round interview with AI questions and a rubric?

Why: consistency across interviewers in different countries depends on a shared rubric, not a free-form chat.

Good answer: one shared rubric and a summary saved to the candidate record, shown live.

Red flag: an unstructured AI chat with no rubric, or a summary that does not reach the record.

7. Is the HRIS integration native or middleware?

Why: a native connector is maintained by the vendor; middleware adds cost and needs an owner on your side.

Good answer: a native connector to your HR system, shown running live.

Red flag: "via our integration partner," with no live demo against your instance.

8. How many weeks until our first market goes live?

Why: a long rollout can cancel out a lower subscription price, and can miss a hiring season.

Good answer: a dated plan with named owners on the vendor side.

Red flag: "it depends," with no timeline or named owner.

9. Where is your support team, and which time zones does it cover?

Why: APAC hiring needs support during APAC working hours, not a ticket queue in another hemisphere.

Good answer: an in-region team covering your working hours.

Red flag: email-only support, or coverage that does not overlap your day.

10. Are AI features billed separately?

Why: AI billed through credits or add-ons can run out during a peak, and is easy to miss in the headline price.

Good answer: AI pricing and usage limits written into the quote.

Red flag: "we can sort out AI pricing later," or no clear overage rate.

How to score the answers

Score each answer as you hear it, not from memory afterward. A simple 1 to 3 scale per question works:

• 3. Proven: shown live on your data, or a specific, verifiable answer.
• 2. Claimed: a credible answer, but described rather than demonstrated. Flag it to verify in a pilot or reference check.
• 1. Weak or evasive: a vague answer, a red flag, or "later."

Add the scores for a rough total, but treat the compliance and data questions (3 and 5) as go/no-go rather than points: a 1 on either is a reason to pause regardless of the total. Capture one line of evidence next to each score while it is fresh, so you can compare vendors a week later without relying on impressions.

After the demo

Two things turn a good demo into a safe decision. Ask for references in your region and industry, ideally customers hiring across the same markets, and ask the vendor to run a short pilot on the live role you brought. A vendor that scored well on claims but will not pilot has told you where the claims stop.

Next step

Bring this list and your own material to every demo, and score the answers in the room. If you are evaluating Moka AI for APAC hiring, bring your resumes and one open role and ask it to answer these ten live.