Hiring a salesperson in Indonesia through an employer of record makes the employment compliant. It does not automatically make your company safe from a permanent establishment finding. Those are two separate exposures with two separate tests, and the second is barely discussed in the market, which is a problem, because sales is the function where it bites hardest. This guide covers how Indonesia actually determines a PE, why revenue roles trip a test that other roles don't, what the December 2025 regulation changed, and the specific role-design, contract and compensation decisions that keep a Jakarta commercial team on the right side of the line.

At a glance: Use compliant employment for the employment problem. Use role design for the PE problem. Keep contract signature, final pricing authority and legal commitment outside Indonesia, document that boundary before the first hire starts, and write the compensation plan so it doesn't quietly reprice your severance liability. Then have Indonesian tax counsel look at the finished structure, not at a template.

The short answer, before the detail

An employer of record arrangement genuinely removes a large category of risk. Your Indonesian salesperson gets a compliant local employment contract. Income tax under PPh 21 is withheld and remitted properly. BPJS Kesehatan and BPJS Ketenagakerjaan are registered and paid. Statutory leave, religious holiday allowance and termination all follow the lawful route. You are not running an unregistered payroll in a jurisdiction where that is exactly the sort of thing that gets noticed.

What it does not do is answer the question Indonesian tax authorities actually ask about a foreign company with commercial activity in Indonesia: is this enterprise carrying on business here in substance? That question looks at conduct. Who negotiates? Who sets the price? Who says yes? Whose name is on the contract, and who effectively decided it would be signed? A salesperson employed by an EOR can still be, in substance, the person through whom a foreign enterprise habitually concludes contracts in Indonesia. Where that is the pattern, the employment paperwork is not the operative fact.

None of that is a reason to avoid Indonesia, or to avoid an EOR. It is a reason to treat the sales seat as a structuring decision rather than a hiring decision, and to be suspicious of anyone who suggests a service subscription makes a tax question disappear.

Counsel check. Whether your arrangement creates a permanent establishment is a question of fact and of the applicable treaty, and it cannot be answered from a description of the employment model alone. Have Indonesian tax counsel assess your own structure before you rely on anything in this section.

How Indonesia decides you have a permanent establishment

Indonesian domestic law defines a permanent establishment (bentuk usaha tetap, usually shortened to BUT) in Article 2(5) of the Income Tax Law (Law No. 36 of 2008). In substance: a form of business used by an individual who does not reside in Indonesia, or who is present for no more than 183 days in a 12-month period, or by an entity not established and not domiciled in Indonesia, to carry on business or activities in Indonesia.

PMK No. 35/PMK.03/2019, issued 1 April 2019, sets out the criteria. Its Article 4(1) gives a three-part test for the ordinary case: there is a place of business in Indonesia; that place is permanent in character; and it is used by the foreign person or entity to carry on business or activities. Article 5(2) then removes the escape route most companies reach for first, the test applies regardless of whether the foreign party owns or leases the place, or has any legal right to use it at all.

Article 5(1) lists what counts as a place of business, and two entries on that list deserve attention from anyone planning a commercial presence: a representative office (kantor perwakilan), and space used for promotion and sales (ruang untuk promosi dan penjualan). Both are named in the regulation itself.

Article 4(2) then lists business forms that are a PE even where the three-part test is not met:

• Construction, installation or assembly projects.

• Provision of services in any form by an employee or any other person, where carried out for more than 60 days within a 12-month period.

• A person or entity acting as an agent whose position is not independent: agen yang kedudukannya tidak bebas. This is the dependent-agent limb, and the one that catches sales. Note what the regulation does not say: it does not require the agent to be exclusive, and it does not make contract-signing authority the test.

• Agents or employees of an insurance company not established or domiciled in Indonesia that collects premiums or bears risk in Indonesia.

And Article 4(3) defines the business or activity broadly: it covers everything done to obtain, collect or maintain income.

The dependent-agent concept is the important one. An agent is treated as dependent where they act for or on behalf of the foreign enterprise, or where their activity is wholly or almost wholly on that enterprise's behalf, and where they do not bear the transaction risk themselves. The mirror-image carve-out matters just as much: a genuinely independent agent, broker or intermediary, acting in the ordinary course of their own business for multiple principals at their own risk, does not create a PE for the foreign enterprise.

Where a tax treaty applies, the treaty definition governs, and the thresholds vary by country. The Singapore-Indonesia agreement, ratified 26 December 2020 and the most relevant one for the many companies that run their Asian operations from Singapore, repays reading closely:

• Article 5(2)(i) creates a PE where an enterprise furnishes services, including consultancy services, through an employee or other person, where the activities continue for periods aggregating more than 90 days within a twelve-month period.

• Article 5(2)(h) sets 183 days for a building site or construction, installation or assembly project. Article 5(4) adds a PE where the enterprise carries on supervisory activities on such a project for more than six months.

• Article 5(3) exempts a fixed place of business maintained solely for storage or display, for purchasing goods, for collecting information, or for advertising, the supply of information, scientific research "or for similar activities which have a preparatory or auxiliary character".

• Article 5(5) is the agency limb. A person acting for or on behalf of the enterprise is deemed a PE if they have, and habitually exercise, a general authority to conclude contracts for or on behalf of it, or if they habitually maintain a stock of goods from which they regularly deliver on its behalf.

• Article 5(7) preserves the independent-agent carve-out for a broker, general commission agent or other independent agent acting in the ordinary course of their own business, but expressly removes it where "the activities of such an agent are devoted wholly or almost wholly on behalf of the enterprise".

That last sentence is the one to sit with, and we return to it below. Check your own treaty rather than assuming: the day counts and the agency wording both differ across Indonesia's network, and this treaty retains a pre-BEPS formulation that several newer treaties do not.

The preparatory-or-auxiliary exemption, and the test that actually decides it. Article 6(1) of PMK 35/2019 provides that, for treaty purposes, a business form meeting the place-of-business criteria but carrying on only preparatory (persiapan) or auxiliary (penunjang) activities is excluded from being a PE. The regulation then defines both: preparatory activities are preliminary activities so that essential and significant activities are ready to be carried out; auxiliary activities are additional activities that facilitate essential and significant ones.

The weight therefore falls on what counts as essential and significant, and Article 6(4) says: activities that are the foreign party's core business; that are an inseparable part of its core business; that directly generate income for it; or that use a significant amount of assets or human resources.

That last limb is the one nobody plans for. An exploration or support presence can lose the exemption on scale alone, without ever generating a rupiah of Indonesian revenue: enough people, or enough assets, and the activity stops being auxiliary as a matter of definition. Article 6(5) adds that the exemption does not apply at all where the foreign party carries on preparatory or auxiliary activities for another party.

One structural point deserves emphasis, because it is the crux of this whole guide:

The services and construction tests are time-based. The agency test is not. There is no 60-day or 90-day grace period on agency PE. It is a conduct test, and a habitual pattern is enough.

Counsel check. The provisions summarised here are set out at a summary level and the treaty that governs your position may differ from the Singapore agreement used as the illustration. Confirm the operative text of your own treaty, and the current text of the regulations cited, with Indonesian tax counsel.

Why sales is the highest-risk role you can put offshore

Most offshore functions are, from a PE perspective, boring in the best sense. An engineer writing code, a designer producing assets, a support agent resolving tickets, a finance analyst reconciling accounts: none of them commits your company to anything. They generate cost, not obligations.

Sales generates obligations. That is the job. And the specific behaviours that make someone good at it are the specific behaviours the agency test describes:

• Negotiating commercial terms. A rep who can move price, scope or payment terms is exercising authority, whether or not the authority is written down.

• Concluding contracts in your name. Under PMK 112/2025, the agency trigger is a dependent agent who lazim menyepakati kontrak atas nama the foreign taxpayer, habitually agrees contracts in its name, and the regulation defines which contracts count: those made in the foreign taxpayer's name, those transferring ownership or granting rights of use over property it owns or controls, and those for the provision of its services. Note what the standard is not: neither this regulation nor the Singapore treaty adopts the post-BEPS "principal role leading to the conclusion of contracts" test. The treaty threshold remains a general authority to conclude contracts, habitually exercised. That is a higher bar than the one applying in many other jurisdictions, which is useful to know, and no reason to test it.

• Holding and delivering from local stock. The classic second limb of agency PE in most treaties. If your rep works from consignment inventory in Jakarta, the conversation is effectively over.

• Acting almost wholly for one principal. The independence carve-out requires genuine independence. A commission-only "agent" whose entire income comes from your company, working your targets in your CRM, is not independent in any meaningful sense.

• Presenting as your local presence. A business card, an email signature, a LinkedIn headline and a Jakarta phone number all reading "Country Manager, Indonesia" are not by themselves determinative. They are evidence, and they are the first thing anyone examining the arrangement will collect.

There is a hard commercial tension here, and it is worth naming plainly rather than pretending it away. Everything that makes a salesperson effective increases PE exposure, and everything that reduces PE exposure slows the sales cycle. A rep who must route every deal abroad for pricing loses deals to a competitor whose local team can say yes. The work is finding the point on that curve you can defend, and documenting why you chose it, not pretending the curve doesn't exist.

Counsel check. Whether a particular role amounts to a dependent agent is decided on the facts of what that person actually does, not on the categories described here. A role that looks safe in the abstract can fail on its conduct, and the reverse is also true.

What changed on 31 December 2025

PMK No. 112 of 2025, Tata Cara Penerapan Persetujuan Penghindaran Pajak Berganda (Procedures for the Application of Double Taxation Avoidance Agreements), was signed on 30 December 2025 and takes effect on the date of its promulgation: 31 December 2025. It was issued to implement Article 50(2) of Government Regulation No. 55 of 2022, and it is the most consequential development in this area in years. Any Indonesian PE analysis produced before it is out of date.

Much of the regulation is treaty administration: the updated DGT forms, beneficial ownership testing, and a principal purpose test that lets the authorities deny treaty benefits where obtaining them was one of the principal purposes of a transaction or arrangement. But Chapter IV is given over to preventing abuse of tax treaties, and it carries three changes that bear directly on how a foreign company structures commercial presence:

• The "preparatory or auxiliary" label stopped being a safe harbour. Whether an activity is excluded now depends on its role in generating income for the foreign enterprise, not on what the arrangement calls it. A "representative office that only does market research", whose staff are in fact working live pipeline, is a question of fact with an obvious answer.

• Time thresholds are aggregated across closely related persons. Contract splitting, breaking one engagement into several so each stays under a day-count threshold, no longer works. PMK 112/2025 adopts the Multilateral Instrument definition of "closely related", and periods run by related parties at the same site are added together. Indonesia ratified the MLI by Presidential Regulation No. 77 of 2019, in force from 1 August 2020 and effective from 1 January 2021.

• The agency and preparatory/auxiliary tests are now written into the treaty-application rules. The regulation sets out when a foreign taxpayer has a PE for treaty purposes, including through a dependent agent who habitually concludes contracts in its name, or who does not conclude contracts but habitually makes deliveries on its behalf, while referring the substantive determination back to the Minister of Finance regulation on determining permanent establishment, which remains PMK No. 35/PMK.03/2019. It also defines the categories of contract that count, so "our rep didn't sign it" is not the end of the analysis where the contract was made in your name.

The practical shift is from mechanical to substantive. Before, PE analysis could be run as arithmetic: count the days, stay under the number, done. After 31 December 2025, the authorities have an explicit mandate to look past the legal form of an arrangement at what is actually happening, and, where they conclude a PE exists, to issue a tax identification number of their own motion, with the obligations that follow.

If your Indonesia structure was designed to sit just inside a threshold, it needs re-examining. If it was designed to reflect what your business actually does, it is probably fine.

Counsel check. This is a recent regulation and implementation guidance from the Directorate General of Taxes is still emerging. Treat the reading here as a summary of the published text, and take advice on how it applies to arrangements you already have in place.

Four ways to build sales in Indonesia, ranked by PE exposure

Every approach trades PE exposure against speed, cost and sales effectiveness. Ranked from highest exposure to lowest:

The detail behind each row, because the ranking is not the whole story:

1. Flying your own people in: highest exposure, lowest apparent commitment

Your existing sales team visits Jakarta on trips. No local hire, no local employment. It feels like the cautious option and it isn't. Your own employees are in-country, furnishing services on behalf of your enterprise, and the day-count clock is running against you: 60 days in 12 months under the domestic rule, or your treaty's threshold. Trips add up faster than anyone tracks, and after PMK 112/2025 you can no longer split the engagement across related entities to keep each one small. This structure also caps your pipeline at whatever a visitor can build, which in Indonesia is not much. This is a relationship market, and quarterly visits do not build relationships.

2. Paying a local commission agent or contractor: high exposure, and the default mistake

You find someone in Jakarta, engage them as an independent contractor, pay commission on closed business. It is fast, it is cheap, it feels asset-light, and it is the worst of the four. Consider what you have built: a person acting for and on behalf of your enterprise, earning almost all their income from you, bearing none of the transaction risk, working your pipeline to your targets. That is a close fit to the dependent-agent description in Indonesian law, and PMK 112/2025 explicitly targets commissionaire and agent structures. You have also created a second, independent problem: misclassification. Indonesian employment law is protective, and a "contractor" who works exclusively for one company under its direction has a credible case for having been an employee all along, with the statutory entitlements that implies.

This is why MixWork does not offer contractor or commission-agent arrangements at all. Not as a limitation of scope, but as a considered position: in a sales context it is the structure most likely to be looked through, and offering it would mean selling clients the exact risk we are supposed to be removing.

3. Employing through an EOR with disciplined role design: materially lower exposure, fast

Your salesperson is a properly employed Indonesian permanent employee, with the employer-side obligations carried by MixWork: compliant contract, PPh 21, BPJS registration and contributions, statutory leave, religious holiday allowance, termination handled through the lawful process. Your exposure narrows to the agency question alone, and the agency question is answered by how you design the role, which is within your control. This is the structure most companies should be using, and the rest of this guide is about doing it properly. The caveat, stated plainly: the EOR handles the employment, you handle the authority boundary. If you give a Jakarta rep the pen, no EOR can help you.

4. Establishing a PT PMA: lowest PE ambiguity, highest cost, slowest

You incorporate an Indonesian foreign-investment company. There is no PE question, because you have a resident taxpayer by design. Your local team can negotiate, price, sign and invoice locally, which is the strongest commercial position available. The cost is real: minimum capital requirements, a licensing process, resident directors, statutory audits, transfer pricing documentation between the PT PMA and the parent, corporate income tax at 22%, and an ongoing compliance load that needs staffing. For a company testing whether Indonesia works, that is a large bet placed before the evidence arrives. For a company with proven Indonesian revenue and a growing team, it is often the right destination. We have written a separate comparison of the PT PMA and EOR routes, including realistic timelines.

The sequence most companies should follow is 3 then 4: prove the market with a compliant employed team and conservative authority limits, then incorporate once the revenue justifies it, by which point you also know exactly which roles you need.

Counsel check. The relative ranking here is general and not a recommendation for any particular business. Each route carries tax, corporate and employment consequences that depend on your sector, your treaty and your commercial model, and the choice should be made with Indonesian tax and corporate counsel.

The Authority Line: nine controls for a Jakarta sales seat

The Authority Line is the boundary between selling and binding. Your Indonesian team can do everything up to it. Crossing it is what creates agency PE exposure. Draw it explicitly, in writing, before the first hire starts: retrofitting it after a tax enquiry opens is worth far less than having it in place beforehand, and contemporaneous documentation is exactly what a substance analysis under PMK 112/2025 will look for.

The nine controls, in rough order of importance:

• 1. Signature stays offshore, and stays real. Every customer contract is executed by an authorised signatory outside Indonesia. "Real" is the operative word, because the regulation reaches contracts made in the name of the foreign taxpayer, not just contracts it physically signed: an offshore signature applied as a formality to a deal the Jakarta rep settled is a weaker fact than it looks. Keep an actual review step with actual discretion, and keep evidence that it is exercised, including deals that came back changed or rejected. The independent-agent carve-out is also worth remembering here: it falls away where activities are devoted wholly or almost wholly to one enterprise, which is true of every employed rep.

• 2. Final pricing and discount authority sits offshore. Publish a rate card the rep can quote from without deviation. Anything outside it (discount, non-standard payment terms, altered scope, custom SLA) routes to a named offshore approver. Log the requests and the decisions. That log is your best evidence that the authority genuinely lives abroad.

• 3. Written authority limits in the employment documentation. The job description and the client-side role definition should state expressly that the employee has no authority to conclude contracts, bind the company, accept orders, or make binding representations on price or delivery. It is not conclusive on its own, conduct beats paperwork, but its absence is conspicuous, and its presence tells the employee where the line is, which is how conduct stays consistent.

• 4. Titles that describe the work honestly. "Country Manager, Indonesia" and "Head of Indonesia" assert local commercial authority. "Business Development Manager", "Account Executive" and "Partnerships Manager" describe a selling role without claiming the power to commit. Apply this consistently across the contract, the email signature, LinkedIn, business cards and any press mention. Inconsistency here is unforced and easily found.

• 5. No local stock, and no local delivery from stock. Habitually maintaining and delivering goods from stock in Indonesia is an independent PE trigger under the agency limb of most treaties. No consignment inventory, no local warehouse, no "we keep a few units at the rep's place".

• 6. No premises that function as your Indonesian office. A lease in your name is the clearest fixed-place-of-business fact there is, and note that Article 5(2) of PMK 35/2019 makes the ownership question irrelevant: a place of business exists "regardless of whether the foreign party owns or leases it, or is legally entitled to use it". A co-working membership in your company's name with your logo at the desk is therefore not the improvement it looks like. What matters is whether the space is available for your unrestricted use and whether you carry on business through it. Where the employee needs professional space (and in Indonesia they usually do, for client meetings if nothing else) take it as workspace provided under the employment arrangement rather than as premises held by the foreign enterprise. MixWork runs its own offices and dedicated workspaces in Jakarta and Singapore, first-party rather than brokered through a third-party operator, which is what makes it possible to structure this cleanly.

• 7. Invoicing and collection stay offshore. Invoices are issued by the foreign entity, in the foreign entity's name, and paid to the foreign entity's account. A rep who invoices locally or collects payment locally is performing the economic function of a resident business.

• 8. CRM stages that record the handoff. Build the authority boundary into the pipeline itself: an explicit stage where the deal leaves the rep and enters offshore approval, with the approver's identity recorded. This turns your compliance position into a byproduct of the sales process instead of a documentation exercise nobody does.

• 9. An annual review of conduct against the documentation. Role boundaries drift, especially with a good rep who is winning. Once a year, sample closed deals and check what actually happened against what the documentation says should have happened. Where they have diverged, either fix the reality or update the structure, but know which one you are doing.

Two things this list does not do, and should not be read as doing. It does not make agency PE impossible: the test is substantive, and a sufficiently central rep can create a PE despite good paperwork. And it does not replace advice on your specific facts, because the analysis depends on your treaty, your product, your sales cycle and the actual behaviour of actual people.

Counsel check. These controls reduce exposure and improve the evidence available to you. They do not eliminate the risk and they are not a safe harbour. Whether they are sufficient in your case depends on your facts, and the design should be reviewed by Indonesian tax counsel before the first hire starts.

The scope in detail: what a Jakarta sales role may and may not do

This is the list to put in the job description, the client-side role definition and the sales playbook, in these words or close to them. The Authority Line is the principle; this is the principle expressed as tasks, which is the only form a salesperson can actually work to. Everything in the first list is ordinary commercial work that does not, by itself, engage the agency limb. Everything in the second list either exercises authority to bind you or creates an independent permanent establishment trigger.

Allowed, the role can do all of this without exercising authority to bind you

• Market and account research. Mapping who buys, how they buy, who signs, what they currently use, what the budget cycle looks like. Under the Singapore-Indonesia treaty, Article 5(3)(d) expressly excludes a fixed place maintained solely for collecting information for the enterprise.

• Prospecting and lead generation in all its forms: outbound calls, email, LinkedIn, referrals, industry associations, conference floors.

• Qualifying opportunities against your own criteria, and disqualifying them.

• Presenting and demonstrating the product or service, including technical deep-dives and proof-of-concept scoping.

• Quoting strictly from a published rate card, with no deviation in price, term, scope or conditions. Quoting a fixed published price is relaying your terms; agreeing a different one is exercising authority.

• Transmitting your standard terms unchanged, and explaining what they mean.

• Gathering and relaying requirements, objections, competitive intelligence and pricing pressure back to the offshore team.

• Building relationships: hospitality, site visits, industry events, the long unstructured relationship work that Indonesian B2B actually runs on.

• Attending negotiations as a participant, presenting your position and taking the counterparty's position back for a decision made elsewhere.

• Coordinating the mechanics of closing: assembling documents, chasing the counterparty's internal approvals, scheduling the signature, explaining the process.

• Post-sale relationship management that does not alter, extend or renew the contract: adoption, satisfaction, escalation, identifying future needs.

• Internal work. CRM hygiene, pipeline reporting, forecasting, account planning, sales enablement.

Not allowed: each of these either binds you or is an independent PE trigger

• Signing or executing any customer document. Contracts, order forms, purchase orders, statements of work, quotations presented as final, letters of intent, and non-disclosure agreements, which people forget still bind the company.

• Agreeing a price or discount outside the published rate card, in any channel, including "I'll see what I can do" followed by a number.

• Agreeing non-standard payment terms, scope, delivery schedule, service levels, warranties, indemnities or governing law.

• Making binding representations on delivery dates, specification or capability: verbally, in a deck, or by email. Email counts, and email is discoverable.

• Accepting an order, however informally. "Send it through and we'll get started" is acceptance.

• Concluding renewals, extensions or expansions. A renewal is a contract. This is the one that most often drifts, because it feels like account management rather than selling.

• Holding stock, inventory, samples or demonstration units in Indonesia, or delivering from them. Article 5(5)(b) of the treaty makes habitually maintaining a stock from which the person regularly delivers an independent trigger, with no reference to contract authority at all.

• Invoicing an Indonesian customer, or collecting payment locally. Not once.

• Occupying premises in your company's name, including a co-working membership taken in your name. Under Article 5(2) of PMK No. 35/PMK.03/2019 a place of business exists regardless of whether you own it, lease it, or have any legal right to use it, so the contract form is not the protection people assume.

• Carrying a title that asserts local commercial authority (Country Manager, Head of Indonesia, General Manager, Director) on the employment contract, business cards, email signature, LinkedIn or in any press mention.

• Being held out as authorised. Anything in your own materials that tells a counterparty this person can commit you is evidence, whatever the employment contract says.

• Working as a commission-only agent bearing no transaction risk. Article 5(7) of the treaty preserves the independent-agent carve-out only for agents acting in the ordinary course of their own business, and expressly withdraws it where activities are devoted wholly or almost wholly to one enterprise. A single-principal commission agent fails that test by design.

The grey zone, stated plainly rather than glossed over

These are not resolved by a list, and any provider who tells you otherwise is guessing. Each needs Indonesian tax counsel on your actual facts:

• The rep who leads the negotiation end to end while offshore approval is a formality. The paperwork says the authority is abroad; the conduct says otherwise, and the analysis is substantive. This is the single most common real-world exposure.

• The recommended discount that is always approved. If the offshore approver has never once said no, the discretion is documentary rather than real. Keep the log that shows it is real, including the refusals.

• Framework agreements and standing purchase arrangements, where the commercial decision was made once and later transactions are mechanical.

• Routine NDAs and mutual confidentiality agreements, which are contracts that bind you even though nobody thinks of them as sales documents.

• A rep who is also the company's only point of contact for an Indonesian regulator, distributor or joint-venture partner, which starts to look like managing local affairs rather than selling.

And the point the lists cannot carry on their own: compliance here is a pattern of conduct, not a document. The scope above only protects you if it describes what actually happens, week after week, and if you can show that from your own records, the pricing-exception log, the CRM approval stage, the deals that came back changed. A perfect role definition with a contrary two-year sales history is worth very little.

Counsel check. This scope is a starting point for discussion with your advisers, not a compliance standard and not advice you should adopt unaltered. The boundaries shift with your product, sales cycle, treaty and the conduct of the individuals involved, and the grey-zone items above are genuinely unresolved. Have Indonesian tax counsel review the scope you intend to use, and reflect it in the employment documentation they approve.

Designing sales compensation that survives Indonesian employment law

This is where offshore sales teams most often create expensive problems, and it is almost entirely absent from generic EOR content, which tends to treat "we run payroll" as the end of the subject. Sales compensation is variable by nature, and Indonesian wage law draws lines that variable pay crosses.

Wage composition is regulated, not free-form. Under PP No. 36 of 2021 on wages, wages may be structured as: wage with no allowances; base wage plus fixed allowances; base wage plus fixed allowances plus variable allowances; or base wage plus variable allowances. Where the structure includes base wage and fixed allowances, the base wage must be at least 75% of the total of base wage plus fixed allowances. You cannot shrink base salary to a token and load the rest into allowances.

Your commission design silently sets your severance liability. Under PP No. 35 of 2021, Article 40(4), the wage used to calculate severance (uang pesangon) and long-service pay (uang penghargaan masa kerja) consists of base wage plus fixed allowances. Where a company's wage structure does not use those components, the calculation runs on wage without allowances; where wages comprise base wage and variable allowances only, the base wage alone is the calculation basis. The consequence for sales roles is direct:

• Genuinely variable commission (paid on results, varying period to period) sits outside the severance base. That is the normal and generally advisable structure for a commercial role.

• A "guaranteed commission", a fixed monthly sales allowance, or a non-contingent draw looks like a fixed allowance, and a fixed allowance is inside the severance base. A guarantee offered casually during a competitive hiring process can therefore raise the cost of every future termination for that employee, permanently.

• The base-versus-variable split is a compliance decision, not only a motivational one. Design it deliberately, document the reasoning, and don't let it be settled by whatever the candidate asks for on the final call.

Termination for missed targets is not straightforward. Indonesia has no at-will employment. Ending an employment relationship follows the grounds and process in PP 35/2021, and performance-based termination generally requires documented, successive written warnings rather than a decision that the number wasn't hit. For a sales team that means the performance management has to be real: written targets agreed at the start of the period, documented reviews during it, and a warning trail if things go wrong. Our guides to severance pay and to employee termination in Indonesia cover the mechanics; the point for compensation design is that you cannot substitute a quota for a process.

The other cost lines, so the model is complete. Employer-side social security in Indonesia runs to roughly 6.24% to 7.74% of salary across the BPJS Ketenagakerjaan programmes, plus the employer share of BPJS Kesehatan, which is subject to its own salary cap. Religious holiday allowance (THR) is a statutory obligation equivalent to one month's wages for employees with twelve months' service, pro-rated below that: budget it as a thirteenth month, not a bonus. And Jakarta's provincial minimum wage for 2026 is Rp 5,729,876, a 6.17% increase set by Governor's Decree No. 1142 of 2025 under PP No. 49 of 2025. That figure matters as a floor and as an indication of wage-inflation direction, but it sits well below the market rate for the MNC-experienced commercial talent that makes an Indonesian sales team worth building. Our BPJS cost guide and PPh 21 employer guide carry the detail, and the cost calculator will give you a landed monthly figure.

Counsel check. Wage structure, severance base and termination process are statutory matters and the treatment of a particular incentive depends on how it is drafted and paid in practice. Have any compensation plan and its documentation reviewed by Indonesian employment counsel before it is offered to a candidate.

Protecting your pipeline: non-compete, non-solicit and confidentiality

A salesperson leaves with relationships. In Indonesia, what you can do about that in advance is more limited than in most Western jurisdictions, and it is worth knowing before you build the plan rather than after.

Indonesian law does not define or regulate non-competition clauses at all, which is where the difficulty starts. The arguments against enforceability are constitutional and statutory: Article 27 of the 1945 Constitution guarantees every citizen work and a livelihood worthy of humanity; Article 31 of Law No. 13 of 2003 on Employment gives employees equal rights and opportunity to choose, obtain and move between jobs; and Article 38(2) of Law No. 39 of 1999 on Human Rights protects free choice of work. On that reading a restraint fails the requirement of a lawful cause under Article 1320 of the Civil Code and is void. The argument the other way is freedom of contract under Article 1338.

The case law has moved toward enforceability, treated as contract rather than employment law. In Supreme Court Decision No. 3549 K/Pdt/2023, decided in late 2023, an elevator-industry company succeeded on appeal against a former employee of ten years who resigned and immediately joined a competitor in breach of a twelve-month non-competition clause. The Court held the employee was obliged to comply with the restrictions in both the employment agreement and the company regulations, that they had acted in bad faith, and that this constituted a breach of contract under Article 1243 of the Civil Code: actionable in the district court. It follows an earlier line running from East Jakarta District Court ruling No. 54/Pdt.G/2017/PN.Jkt.Tim, upheld on appeal in 2018 and by the Supreme Court in 2019.

Two qualifications that matter more than the headline. First, Indonesia does not apply stare decisis. Courts are not bound by precedent and each case is decided on its own facts, so these decisions show what is possible, not what is settled. Second, the considerations the courts actually weighed are the drafting brief: the restraint must be fair and proportionate to the protection sought, a senior employee with access to confidential information can bear more restriction than a general one; duration and scope must be clear, specific and not excessive; the employer's investment in the employee (training, courses, exposure to proprietary market information) supports reasonableness and helps prove the access existed.

And the practical point that decides whether any of it is worth having: put in a liquidated damages clause. Indonesian courts require a claimant to prove loss with a detailed, concrete calculation, and the value of misused confidential information is close to impossible to compute: particularly where, as with most Indonesian private companies, revenue figures are not published. Without a pre-agreed damages figure, you can win the point and recover nothing.

What to rely on instead, in priority order:

• Confidentiality obligations over customer data, pricing and pipeline. More robust, and more likely to be enforced, than a restraint on employment itself.

• Customer non-solicitation, narrowly scoped. A twelve-month restriction on soliciting named accounts the employee actually worked is far easier to defend than a blanket prohibition on working for competitors.

• Operational controls that don't depend on a court. Your CRM is the system of record, not the rep's phone. Relationships are multi-threaded, with a second contact point on every meaningful account. Handover procedures are written before they are needed. Notice periods are long enough to run a real transition.

• Retention, which is the only genuinely reliable protection. A rep who stays doesn't take the pipeline anywhere. MixWork's twelve-month retention runs above 90%, and it is not an accident: a dedicated HR manager backed by a full HR team, monthly check-in calls with both the employee and the client, professional workspace, benefits, and regular professional learning sessions covering the AI and software tooling the role actually uses. For a sales team, where the cost of losing someone is measured in lost pipeline rather than replacement fees, that is the control that matters most.

Have any restraint clause drafted and reviewed by Indonesian counsel. This is not a place for a template borrowed from another jurisdiction.

Counsel check. Indonesia does not apply binding precedent, so the decisions referred to here show what is possible rather than what is settled, and outcomes remain case-specific. Any restraint, non-solicitation or confidentiality clause should be drafted and reviewed by Indonesian counsel rather than adapted from another jurisdiction.

Why Jakarta, and how Indonesia compares across Southeast Asia

The PE analysis above is a cost of doing something worth doing. The reason to accept it:

Indonesia is the market, not a proxy for it. With a population of roughly 287.9 million, Indonesia is Southeast Asia's largest consumer market and the world's fourth most populous country. BPS-Statistics Indonesia recorded growth of 5.61% year-on-year in Q1 2026, and the OECD's June 2026 outlook has the full year at 4.7%. Growth is led by household consumption, which is more than half of output by expenditure: precisely the shape of economy where a local commercial team compounds rather than merely covers.

Jakarta is where the buying happens. The major banks, the conglomerates, the regional offices of multinationals, the government ministries and the technology companies are concentrated there. For an enterprise sales motion it is close to a prerequisite. Indonesian B2B buying runs on relationships built in person, and that is not a preference a foreign vendor gets to overrule.

On talent, the argument is composition rather than cost alone. The professionals MixWork places average around six years' experience, are typically in their late twenties, come from multinational or corporate backgrounds, and were educated at Indonesia's leading universities. Universitas Indonesia, Gadjah Mada, Institut Teknologi Bandung, Airlangga and their peers, all of which sit inside the QS World University Rankings 2026. Six years is the band where someone delivers independently without yet carrying management overhead, and for a commercial role that is the highest-value point on the curve. Six years of MNC-standard process is a different proposition from six years of experience in general, and it costs a fraction of the equivalent hire in Singapore or Sydney.

English is rarely the obstacle buyers expect. Proficiency is very high among Jakarta's professional class, and near-native among the professionals MixWork places: they write to clients, present to stakeholders and run meetings without an intermediary, and nobody reviews their emails before they go out. For a sales role that is table stakes, and it is met.

On AI-augmented commercial work, Indonesia is ahead of the global average. Microsoft's Work Trend Index 2026, released 30 June 2026, found 33% of Indonesian workers qualify as "Frontier Professionals", advanced AI users, against 16% globally; 93% treat AI output as a starting point rather than a final answer, against 86% globally; and 62% name critical thinking as their priority skill, against 46% globally. That maps onto what we see in our own placements, where AI tooling is part of daily practice across every function and adoption has moved noticeably in the six months to August 2026. For a sales team, that is the difference between reps who research, personalise and follow up at volume and reps who work a list.

How the region compares for a sales hire, honestly:

On that last column, the honest detail. Indonesia's genuine disadvantage is employment protection. Severance and long-service entitlements under PP 35/2021 are more protective, and therefore more expensive, than Singapore's or Malaysia's, and termination is a process rather than a decision. That is a real cost and we are not going to pretend otherwise. What changes the arithmetic is that the cost is a function of turnover, and above 90% retention, with performance managed and documented from month one, it is a liability you rarely realise. Handled properly it is a known, budgetable number. Handled badly it is a surprise.

The conclusion, stated as a rule: if the revenue you are chasing is in Indonesia or flows through it, the sales seat belongs in Jakarta, and the only question is how you structure it. Indonesia's employment protections are a reason to bring a competent partner, not a reason to sell from somewhere else.

Counsel check. The severance and termination position described here is a summary of the statutory framework and the actual cost of an individual exit depends on tenure, wage structure and the grounds relied on. Model it with Indonesian employment counsel rather than from the general position.

What MixWork does that a platform subscription doesn't

Three differences that matter specifically for a sales build:

We find the person, not just employ them. Most EOR providers require you to arrive with a candidate already selected, the employment is the product and sourcing is your problem. MixWork's experienced recruitment team sources, screens, verifies and background-checks, from 10% of first-year salary, and then employs the person on the same flat EOR fee as anyone else. For a sales hire this matters more than for most roles: assessing Indonesian commercial talent from abroad is hard, the signals that transfer from Western hiring are unreliable, and the cost of a bad first sales hire in a new market is a lost year rather than a lost fee.

Total Care 360 is included, not metered. A dedicated HR manager backed by a full HR team, monthly check-in calls with the employee and separately with you, engagement and dispute resolution, performance and attendance monitoring, and regular professional learning sessions on current AI and software tooling. The platform model prices the equivalents as separate modules brokered through third parties; ours are first-party and in the flat fee, from USD $249 per employee per month. For a sales seat, the monthly employee call is the mechanism that surfaces a motivation problem in week three rather than in the quarterly number.

We are physically in the region, and first-party about it. Offices and dedicated workspaces in Singapore and Jakarta run by MixWork rather than resold from a co-working operator, from $199 per workspace per month, and Managed IT from $99 per device per month with devices held in-region, no deposit and no upfront payment. When a client meeting needs a room, or a laptop dies the day before a pitch, the accountability doesn't route through a vendor's vendor.

And one thing we deliberately don't do: contractor and commission-agent arrangements. Full-time permanent compliant employment only. In a sales context, that constraint is the product.

A 90-day build plan

Days 1, 15. Structure before hiring. Define the Authority Line for the role and write it down: what the rep can quote, discount and commit, and where each of those stops. Confirm which tax treaty applies to you and its actual thresholds. Have Indonesian tax counsel review the intended structure (role definition, authority limits, contracting and invoicing flow) while it is still cheap to change. Decide the base-versus-variable compensation split with the severance-base consequence in front of you, not after.

Days 10, 45. Sourcing in parallel. Brief the search on the role as you have defined it, including the authority limits, because they change which candidates fit: you are looking for someone who can build genuine relationships and run a complex sale without needing the pen, which is a specific profile. Interview for that explicitly. Verify employment history and references properly, this is a market where that work pays for itself.

Days 30, 60. Employ and equip. Compliant permanent employment contract with the authority limits documented. PPh 21, BPJS Kesehatan and BPJS Ketenagakerjaan registered from day one. Workspace and equipment in place before the start date. CRM configured with the offshore approval stage built in, so the boundary is enforced by the process rather than remembered by the person.

Days 60, 90. Operate and evidence. Written targets agreed and documented. First monthly HR check-in inside week four. Log every pricing exception request and its offshore decision from the first deal onward. By day 90 you should have a working pipeline and a contemporaneous evidence trail showing where commercial authority actually sits, which is the thing that has value if the arrangement is ever examined.

Questions to ask any provider before you sign

Answers to these separate the providers who understand this problem from the ones selling a subscription:

• "How does hiring through you change our permanent establishment position, and what does it not change?" A provider claiming to eliminate PE risk for a sales role is either not thinking about it or is willing to tell you something untrue. The correct answer distinguishes employment exposure from agency exposure.

• "Will you review our role design and tell us if it is creating exposure?" Or is your responsibility strictly the payroll and the contract, with everything upstream being our problem?

• "Can you support us on the ground in Jakarta: meetings, workspace, equipment, someone who is actually there?" Or is this administered remotely by a support queue in another timezone?

• "Do you source candidates, or do we arrive with one?" And if you do, what does it cost on top?

• "Who reviews the employment contract's authority limits, and can we get Indonesian counsel input on the structure?"

• "What is included in the monthly fee, itemised, and what is an add-on?" Compare the total, not the headline. The category norm is a low headline with the substance metered separately.

• "What is your retention rate, over what period, and measured how?" For a sales team this is the number that determines whether Indonesian severance protection is a budget line or a shock.

The bottom line

Building a sales team in Indonesia is one of the better commercial decisions available to a company with Southeast Asian ambitions, and Jakarta is where it should sit. The compliance work is real but tractable, and it divides cleanly into two halves that get confused constantly.

The employment half is a solved problem. A compliant permanent contract, PPh 21, BPJS, statutory leave, THR, lawful termination, that is what an employer of record is for, and a competent one removes the exposure.

The permanent establishment half is a design problem, and it is yours. It turns on authority: who quotes, who discounts, who signs, who binds. No service subscription answers that for you, and after PMK 112/2025 the answer has to be true in substance rather than merely documented. Draw the Authority Line deliberately, build it into the sales process rather than into a policy nobody reads, evidence it from the first deal, and have Indonesian counsel look at your specific facts.

MixWork does the employment half completely, sources the person, and supervises the role on the ground in Jakarta. On the PE half we will tell you what we see and where we think the line should sit, including when the problem is your role design rather than your paperwork, which is the answer nobody selling a platform has a reason to give you.

Talk to us about the role before you post it. That is the conversation where this gets cheap.

This guide summarises Indonesian tax and employment regulations as at 20 August 2026, including Article 2(5) of Law No. 36 of 2008, PMK No. 35/PMK.03/2019, PMK No. 112 of 2025, PP No. 35 of 2021, PP No. 36 of 2021, PP No. 49 of 2025 and the Singapore-Indonesia tax treaty. It is general information, not legal or tax advice. Permanent establishment determinations depend on the specific facts of each arrangement and on the applicable tax treaty, and regulations and thresholds change. Before acting, obtain advice from qualified Indonesian tax and employment counsel on your own circumstances.