A New Country, a New Set of HR Questions: The First Ninety Days of Support

TL;DR

• A new market generates far more HR questions per person than an established one, and the mix is different, so planning from your company average understates both the volume and the difficulty.
• If you are hiring into a country where you already have a population, you do not need this. The content exists and the questions are answered.
• Three waves arrive in a predictable order: setup and access in the first fortnight, pay and benefits around the first full cycle, then policy and entitlement questions once people start planning.
• Most of the wave-one questions are universal and answerable from existing content. Most of the wave-three questions are not, and they are the ones nobody has written anything for.
• Write the answers before the first joiner starts, not after the questions arrive. Twelve documents covers most of it.
• The useful measure is questions per head against your company average, tracked monthly. It should start at several times the average and fall. If it does not fall, something is missing.

The Market That Generated More Tickets Than a Region

A company opened in a new country with four people. The HR team planned for it sensibly, in the sense that nobody ignored it: there was an onboarding checklist, the handbook was shared, and somebody was nominated as the point of contact.

In the first quarter those four people generated more support questions than a 60-person established region. Not because anything had gone wrong, and not because they were unusually demanding. They asked about access to systems nobody had provisioned for that country, about a payslip in a format they had not seen before, about whether a public holiday they expected was recognised, about which of two leave policies applied to them, and about a benefits election with a deadline they had been told about by a system that did not know where they worked.

Every one of those was a reasonable question with no written answer. The planning error was not underestimating the number of people. It was assuming that question volume scales with headcount, when in a new market it scales with the number of unanswered firsts, and a market of four people has almost exactly as many firsts as a market of forty. The fixed cost of a new market, in support terms, is nearly independent of its size.

Recommended Tools

Best tools for HR Operations

B BambooHR HR management system R Rippling HR & IT automation W Workday Enterprise HRIS

View all recommended tools

This is what the first ninety days actually look like, and most of it can be written down in advance.

When You Don't Need to Plan for This

When the manual way is genuinely fine

You are hiring into a country where you already employ people. The firsts have been encountered, the answers exist somewhere, and the new joiner's experience is an ordinary onboarding. The only thing worth checking is whether those answers were ever written down or whether they live in one person's memory, because the second is a risk you will discover when that person is on leave.

When friction starts appearing

The nominated point of contact starts forwarding questions they cannot answer, or answering them by asking somebody else and relaying the reply. That relay pattern is the signal, and it is worth noticing early because the relay is doing real work and producing no written artefact, so the next joiner will generate the same questions again.

When it becomes a liability

The point where a new joiner has missed something with a deadline, most commonly a benefits election or a payroll cut-off, because the communication about it assumed a context they did not have. The exposure is sharper in a new market than anywhere else, since the employee had no colleague to check with and no previous cycle to pattern-match against.

The edge case that forces it

Opening two markets at once, or opening a market with a cohort rather than an individual. A cohort of eight starting on the same day produces all three waves simultaneously rather than in sequence, which removes the natural spacing that makes the single-joiner case manageable.

Five Questions People Ask First

"How much volume should we expect?" Several times your company average per head for the first two or three months, falling towards normal by about month six if the content work gets done. The number that matters is not the absolute volume, which is small, but the ratio, because a market stuck at three times the average in month eight is telling you that the answers still do not exist.

"Can we not just share the handbook?" You can, and it answers the universal questions, which is a real contribution. It will not answer the wave-three questions, and worse, it will appear to answer them, because an employee reading a confident paragraph written for another country has no way to know it was not written for them.

"Who should be the point of contact?" Somebody with time and an escalation route, and ideally somebody in or near the market. The common failure is nominating a senior person in head office whose availability is poor, which converts every question into a delay. A junior person with a clear route onwards serves people better than a senior one who is in meetings.

"When should the content be written?" Before the first start date. This sounds obvious and almost nobody does it, because the content work has no deadline attached while the hiring does. Attaching it to the offer stage rather than the start date is the practical fix.

"Do we need local advice?" For the substance of anything that differs by jurisdiction, yes, and the useful thing is to get it once, early, covering the predictable list rather than question by question as they arrive. Answering these reactively is slower, more expensive and produces inconsistent answers across the first few joiners.

The Three Waves, and What Each Needs

Wave one, the first fortnight: access and setup

Logins, equipment, expense submission, who to contact, how to book time off mechanically, where the handbook is, how to read a payslip layout they have not seen.

Almost all universal, almost all already answered somewhere in your content, and the failure is findability rather than absence. This wave is also the easiest to over-serve: a well-made starter page covering twelve things removes most of it, and it is reusable for every subsequent market without change.

The one genuinely local element is provisioning. Systems that are configured per country, device shipping, and anything requiring an account with a local provider. These are the questions where the honest answer is often that somebody is working on it, and saying that explicitly beats silence.

Wave two, the first full pay cycle: pay and benefits

The first payslip generates questions in every market, and in a new one it generates them from everybody at once. Format, deductions they do not recognise, timing, and the mechanics of any benefit enrolment.

Partly universal and partly local, and this is the wave where the deadline risk concentrates. Benefits windows and payroll cut-offs are dated, and a new joiner has no previous cycle to anchor on. Everything dated should be communicated explicitly with the date, the time and the timezone, rather than relying on a system notification that may not know where the person works.

Wave three, month two onwards: policy and entitlement

Once people stop setting up and start planning, the questions change character. Leave entitlement and carry-over, notice arrangements, probation, sick pay, parental arrangements, whether a particular public holiday applies, how performance review timing works for a mid-year joiner.

This is the wave with almost no existing content, the highest proportion of jurisdiction-specific answers, and the greatest potential to go wrong. The substance of most of these differs by country and by individual circumstance, which means the honest content is frequently a short section saying so and naming who handles it, with the specifics coming from local advice rather than from a template.

It is also the wave that arrives after everybody has declared the market launch finished.

The Twelve Documents That Cover Most of It

Written once per market, before the first start date, and most are short.

Four that are universal and reusable across every market: the starter page for systems and access, how to submit expenses, how to book leave mechanically, and who to contact for what, with names.

Four that are local and factual: the public holiday list for the year, the payroll calendar with cut-off dates and timezones, the benefits enrolment window with dates, and a plain-language explanation of the local payslip layout.

Four that are local and must be written carefully, because the substance differs by jurisdiction: leave entitlement and carry-over, probation arrangements, sick pay and absence reporting, and notice. For each of these the content should state the process that is consistent, say plainly that the specifics depend on where somebody is employed and on their circumstances, and name the person who handles it. The substantive position for each should come from local advice rather than from inference, and the document's job is to route people reliably rather than to assert an answer.

That is twelve documents, of which four are reusable, four are lists of dates, and four are short routing sections. A day of work, plus whatever time the local advice takes to obtain, and it removes most of the ninety-day load before it arrives.

Writing the Four Reusable Documents Once

The four universal documents are the best return in this whole article, because they are written once and then serve every market you ever open. So it is worth building them deliberately as templates rather than as content for the first market that happened to need them.

The systems and access starter page. Twelve items, each one line, each linking to a how-to rather than containing one. The discipline is to list what a new person needs in their first three days and nothing else, because a comprehensive page gets skimmed and a twelve-item page gets read. Keep the country-specific provisioning items in a short separate block at the bottom, which is the only part that changes per market.

How to submit expenses. Reliably one of the highest-volume early questions, and almost entirely universal. The one local element is currency and any local approval step, so write the process generically and leave a single marked line for the local detail.

How to book leave, mechanically. Note the word mechanically. This document covers which system, which buttons, who approves and how far ahead, and it says nothing whatsoever about entitlement, which is jurisdiction-specific and belongs in one of the careful four. Separating the mechanism from the entitlement is what keeps this document reusable, and conflating them is the most common reason a leave document cannot be reused.

Who to contact for what, with names. The shortest document and the one that prevents most escalation failures. Four or five rows: this kind of question goes to this person, with a named backup. Review it when anybody changes role, because this is the document that rots fastest and the one whose rot is most visible to a new joiner.

Build all four with a marked placeholder for the local block, so opening market number four is a matter of filling four short blocks rather than writing four documents. After three markets the whole ninety-day content job should be an afternoon, and most of that afternoon is checking dates.

One structural choice worth making early: keep these four in the same place as the rest of your employee content rather than in an onboarding system. Onboarding tools tend to hold content that is only visible during onboarding, and every one of these four documents is something people come back to months later, particularly the contacts list and the expenses process. Content that disappears after week two generates a second round of the same questions in month five, from the same people, who now remember that the answer existed and cannot find it.

And date the four local blocks visibly, with the year on the holiday list and the payroll calendar. These are the documents most likely to be read a year later by somebody who assumes they are current, and a visible year is the cheapest protection against that.

How to Choose: Five Questions Before You Talk to Any Vendor

Can content be scoped to a population? The whole problem in wave three is general content being served to people it was not written for, so the ability to mark content as applying to a specific market, in a way the tool respects, is the central capability.

Does it answer from your documents, with the source visible? In a new market you will be correcting content weekly for the first two months. A document-trained tool makes that a content edit. A tool generating from general knowledge does not.

Can it refuse and route by topic? Wave three is mostly questions that should route rather than be answered, and a tool that answers them confidently in a market where you have no local content is the worst case in this whole article.

Does it report volume by the asker's market? This is how you know whether the ratio is falling. Without it you are reading a company aggregate in which four people are invisible.

What does adding a market cost? On a per-seat basis the local point of contact is a licence, and on a flat or per-workspace basis they are not, which is a small number that recurs with every market.

What the Tools Publish

Each figure below was read from the vendor's own page on 5 and 7 October 2026, and each vendor is described on its own, because a paragraph mixing several vendors and several numbers is how a real price ends up attached to the wrong product.

Matram

Disclosure: Matram is owned by the same people who publish HROpsLab. It appears here because it competes in this category and is assessed against the same criteria as everything else on this page, with its limitations stated in the same detail.

Two properties fit this use. It answers from your own documents, so the weekly content corrections of a new market launch are edits rather than retraining, and it publishes 95+ languages, which matters when the new market's working language is not yours. It prices flat at $29, $69 and $199 a month with seats unlimited, so each new market's local point of contact does not add a licence line, and there is a 30-day trial with no card required.

Where it struggles: no free tier, so it cannot be parked on a four-person market indefinitely at zero cost. It answers rather than actioning tickets, which is a real limit in wave one where much of what people need is provisioning rather than information. And it is not an HR workflow platform.

SiteGPT

Document-trained and publishes the same 95+ languages, so neither leads on reach. Priced at $468 and $948 billed yearly, with the monthly-equivalent figures applying only on that annual commitment. Starter covers one chatbot and a capped page count, which is a constraint if you wanted a separate deployment per market.

Chatling

Document-trained with a free tier, which makes it a cheap way to pilot content for a single small market before committing. Publishes 80+ in one homepage section and over 85 in another.

Botsonic

Publishes 50+ languages, plainly stated. Check your new market's language against it specifically rather than assuming.

Crisp

A broader platform with live chat, priced per workspace at Free, $45, $95 and $295 a month, so adding a point of contact per market costs nothing. The live chat element suits wave one, where a short real-time exchange often resolves a provisioning question faster than any document. No published language count.

Zendesk

Suite Team is $55 per agent per month, with the Copilot add-on a further $50 per agent per month. Strong workflow, which matters for the provisioning half of wave one, and the per-seat basis means each market's contact is a licence. No published language figure, and the pricing page geo-redirects.

Freshservice

Tiered per agent at $19, $49 and $99 per month, with Freddy AI priced separately at $29 per agent per month. Covers workflow as well as answers, so it can action some provisioning rather than only explaining it. No published language count.

Three vendors publish no price at all. Leena AI and Moveworks are demo-only, and Moveworks has no working pricing page. Tidio has replaced plan pricing with a usage calculator, so its on-page figures are conversation counts rather than money.

The Comparison

The Decision Table

The Questions Nobody Asks, Which Are the Problem

Volume is the visible half. The questions that do not arrive are the other half, and in a new market they are more dangerous than anywhere else, because a new joiner has no colleague to ask instead and no previous cycle to pattern-match against.

Three things routinely go unasked. The first is anything the employee assumes is the same as at their previous employer, which they have no reason to question. Payroll timing is the classic: somebody who has always been paid on the last working day will not ask when payday is, and will discover the answer when it matters.

The second is anything they assume somebody is already handling. Provisioning sits here. A new joiner missing access to a system they know exists will frequently wait several days before raising it, on the reasonable assumption that it is in progress, which it may not be.

The third is anything they think they should already know. This is the largest and the hardest to address, and it falls heaviest on junior joiners and on anybody working in their second language. The cost of asking feels higher to somebody who suspects the answer is obvious, and in a market of four people there is nobody to check with first.

The remedy for all three is the same and it is not a document. It is a scheduled conversation: fifteen minutes in week one and fifteen minutes after the first pay cycle, with somebody whose explicit job in that conversation is to volunteer information rather than to answer questions. Give them a short list to run through, including payday, the benefits window, the holiday list and who to contact, and the pattern of unasked questions largely resolves itself.

But do it twice rather than once. The week-one conversation lands while the person is saturated and remembers little of it, and the post-payroll conversation lands when the first real questions have formed, which is why the second one surfaces more than the first.

Measuring the Ninety Days

Three numbers, checked monthly rather than quarterly, because the whole period is one quarter long.

Questions per head against the company average. Expect three to five times the average in month one. The shape matters more than the level: it should fall each month. A flat ratio means the questions are recurring because the answers are still not written, and a ratio that falls sharply to well below average means the population has disengaged rather than been satisfied, which is the same trap as any other market.

Repeat topics. The list of topics asked more than twice in the period is your content backlog, ordered for you by the people who needed it. This is the most useful artefact the first ninety days produces and it is routinely discarded.

Questions the contact could not answer. Track these specifically, because they are the jurisdiction-specific set and they are the input to your local advice conversation. Batching them is considerably cheaper than asking one at a time, and it produces consistent answers rather than a sequence of individual judgements.

A fourth, after the first market: how much of the content you wrote was reusable. The answer tells you what to build as a template for the next market, and over three or four markets it turns a day of work per market into an afternoon.

Two cautions about reading these numbers in a market of four or six people. Nothing here is statistically meaningful at that size, and treating it as though it were produces over-reaction in both directions: one quiet month reads as success and one busy week reads as a crisis. Use the figures to decide what to ask about rather than to conclude anything, and lean on the fifteen-minute conversations for the actual signal, because at this scale talking to everybody is genuinely possible and is more informative than any dashboard.

The second caution is about attribution. A high ratio in month one is normal and healthy, and a team that is measured on reducing it will start discouraging questions, which is the opposite of what you want in the one period where asking is the behaviour you most need. If this gets reported upwards at all, report it with the expected curve attached, so that a high number in month one is read as the plan working rather than as a problem to suppress.

What Changes When the Market's Language Is Not Yours

Everything above assumes the new market works in a language your content already exists in. When it does not, the ninety days get harder in three specific ways, and the order of operations matters.

Scope before you translate. The instinct is to translate the handbook for the new market, and doing that first is a mistake, because you will faithfully translate paragraphs that were written for another country and hand a new population confident answers that do not apply to them. Mark which sections apply to whom first, then translate, or you have multiplied an existing problem into a second language.

Translate the four reusable documents and the four date lists first. These are high-volume, universal or purely factual, and they absorb most of wave one and wave two. They are also safe to translate with ordinary care, because they contain process and dates rather than terms of art.

Handle the careful four differently. Leave, probation, sick pay and notice are exactly the content where a smooth translation of a general paragraph is most harmful. Since the honest version of each is a short routing section rather than an answer, the translation job is small, and it should be done properly, with the named contact and the stated timeframe checked in the translated version. A name rendered phonetically or a team name translated descriptively turns a named route back into a generic one.

So the sequence is scope, then translate the eight safe documents, then write and carefully translate the four routing sections. Doing it in that order costs no more than doing it badly and removes the worst available outcome, which is a new population in a new country receiving fluent, confident, inapplicable answers about entitlement in their first month.

One practical note on volume. A new market in a new language will sit at the top of the three-to-five times range rather than the bottom, and it will take longer to fall, because comprehension questions stack on top of the firsts. Plan the ratio accordingly rather than concluding at month four that something has gone wrong.

What Getting This Wrong Costs

The direct cost is a missed deadline, usually a benefits election, and it falls on somebody in their first weeks who had the least ability to see it coming.

The second cost is the first impression, which is harder to repair than it is to get right. A new joiner in a new market forms a view of the company's competence from exactly these interactions, because they have almost no other evidence. Five unanswered questions in the first fortnight is a specific and lasting impression, and it is formed by people who are also your only advocates in that market for the next hiring round.

The third cost is the repetition. Every question answered verbally and not written down will be asked again by the next joiner, and the one after. A market that has had six joiners and no written content has answered the same twelve questions six times, which is both expensive and a worse experience each time, because the person answering is increasingly sure the answer is obvious.

The fourth cost lands on the hiring itself, which is the argument that tends to get this funded. A market's first few employees are the people who will refer candidates, speak about the company to their network and anchor the local culture, and their view of whether the organisation is competent is formed almost entirely from administrative interactions in their first two months, because there is nothing else for them to judge by yet. A new market that is hard to join is harder to hire into eighteen months later, and that effect is slow, invisible in any support metric and expensive in a way that no content backlog captures.

So the question worth asking before a market opens is not who the point of contact will be. It is which twelve things are written down.

When You're Ready to Move Beyond Answering as They Come

Nearly every first market is handled reactively, and it usually works out, because the first joiners are often senior, motivated and willing to chase. That is a genuine reason it does not feel like a problem: the people who experience the worst version of it are the most capable of absorbing it.

What makes it stop working is the second and third joiner, who are less senior, less connected and have less patience for an arrangement that depends on knowing who to ask. They also arrive after the launch has been declared complete and attention has moved on, which is why the second cohort in a new market is frequently served worse than the first.

The sequence is short and almost entirely writing. Attach the content work to the offer stage rather than the start date, so it has a deadline. Write the four reusable documents once, for every market you will ever open. Produce the four local lists of dates, which are quick and prevent the deadline failures. Get local advice once on the predictable jurisdiction-specific list rather than question by question, and turn it into four short routing sections rather than four confident answers. Then track questions per head monthly and treat a ratio that stops falling as a content backlog rather than as a quirk of the market.

Frequently Asked Questions

HROpsLab takes no vendor money and publishes no paid placements, and does not state what any jurisdiction requires, which is why the wave-three topics above point at local advice rather than at an answer.