Arbitration: A Comprehensive Guide to Alternative Dispute Resolution

In brief. Arbitration lets parties resolve disputes privately, before decision-makers of their own choosing, and—through the New York Convention—enforce the result in most of the world. Domestically, it runs on the Federal Arbitration Act and a powerful federal policy favoring arbitration; internationally, it runs on the New York Convention and the UNCITRAL Model Law. That power rests on a sophisticated framework of treaties, statutes, institutional rules, and case law. This guide walks through the whole arc, from drafting the clause to enforcing the award, using a single cross-border dispute as a worked example. It reflects the law as of early 2026, including the UK Arbitration Act 2025. It is not legal advice; rules and seats differ, so consult qualified counsel on any specific matter.

Every commercial relationship carries within it the seed of a dispute. Goods arrive late or damaged. Software fails to perform as specified. A joint-venture partner quietly diverts opportunities. When these conflicts arise—and they will—the question that matters most is not whether the parties have a remedy, but where and how that remedy will be pursued. For a growing share of businesses worldwide, the answer is arbitration.

Arbitration is, at its simplest, an agreement to resolve disputes privately, before one or more decision-makers the parties choose, rather than in the courts of any particular country. That simplicity is deceptive. Beneath it lies a framework spanning international treaties, national legislation, institutional rules, and decades of judicial interpretation. Mastering that framework is the difference between a practitioner who files an arbitration and one who wins it—and, just as often, the difference between a business that drafts a clause it can rely on and one that discovers, too late, that its clause was a trap.

To keep the discussion grounded, this guide follows one hypothetical dispute. Helvex Turbines AG, a Swiss manufacturer of industrial gas turbines, agreed to supply six turbines to Pacate Power Pte. Ltd., a Singapore-based developer building a power plant in Southeast Asia. Their supply contract is governed by Swiss law but contains an arbitration clause seated in Singapore under institutional rules. When the turbines allegedly underperform, Pacate withholds the final payment of US$18 million and accuses Helvex of misrepresenting the equipment's efficiency; Helvex counters that Pacate operated the turbines improperly. As we move through each stage, we will watch how the doctrine plays out for Helvex and Pacate—because every rule looks different depending on whether you are the party demanding arbitration or the one resisting it. (This scenario is hypothetical and illustrative only.)

This guide is the definitive treatment of how arbitration works. It pairs naturally with two companion pieces that approach the subject from different angles. If your question is which forum to choose—court, arbitration, or mediation—and how those roads compare, start with Arbitration, Mediation, and Choosing a Dispute-Resolution Forum. If your question is the U.S. institution most parties will actually encounter, see our detailed analysis of AAA commercial arbitration. And before you commit any dispute to any forum, the discipline of evaluating and assessing a civil lawsuit—quantifying exposure, mapping the proof, and working backward from the goal—applies with equal force in the arbitral setting. This article is the connective tissue: the deep mechanics that those guides assume.

Where Arbitration Sits: The ADR Spectrum

Arbitration is one species of a larger genus called alternative dispute resolution (ADR), and it helps to fix its place on the spectrum before diving into the mechanics. At one pole sits negotiation, in which the parties bargain directly with no neutral involved and no one but the parties can impose a result. A step toward formality is mediation, in which a neutral third party helps the parties reach a settlement they design themselves; the mediator persuades and shuttles but cannot decide. At the far pole sits adjudication—a binding decision imposed by a neutral—of which litigation and arbitration are the two great branches. Arbitration is private adjudication: the parties hand a neutral the power to decide, but they choose the neutral and design the process.

That single distinction—between facilitating a settlement and deciding the dispute—is the line that separates mediation from arbitration, and confusing the two is a surprisingly common drafting error. A mediator who "rules" has exceeded the role; an arbitrator who merely "suggests" has shirked it. The two can be combined (the much-debated "med-arb" model, where the same or a different neutral mediates first and arbitrates what remains), but the roles must be kept analytically distinct. Mediation in the United States is governed less by a unified statute than by a patchwork of court rules and the Uniform Mediation Act (2001), adopted in roughly a dozen states, whose central contribution is a robust mediation privilege protecting the confidentiality of what is said in the room. We treat mediation and the forum-choice question at length in our companion guide to choosing a dispute-resolution forum; here, our subject is the adjudicative branch—arbitration—and we will mention mediation mainly where the two intersect, as in the multi-tiered "escalation" clauses that require negotiation, then mediation, then arbitration in sequence.

Why Arbitration? The Case for Leaving the Courthouse

To appreciate arbitration, it helps to understand the problem it solves. Litigation is a creature of sovereignty. Courts derive authority from the state, apply the state's procedural rules, and run on the state's calendar. For a purely domestic dispute between parties in the same jurisdiction, that system works tolerably well, if slowly. Introduce a cross-border element, and its limits become acute.

Our hypothetical illustrates the difficulty. If Helvex and Pacate had no arbitration clause and a dispute arose, whose courts would hear it—Swiss, Singaporean, or the courts of the plant's location? In what language? Over how many years and appeals? And if Helvex won a Swiss judgment, would a Singaporean or third-country court enforce it against Pacate's assets? The answers are uncertain, jurisdiction-dependent, and frequently unsatisfying. Arbitration solves each problem through a single mechanism: consent. Because arbitration rests on the parties' agreement rather than any state's coercive authority, it lets them design a process that is genuinely neutral—as to forum, language, and decision-maker—and, critically, enforceable almost anywhere through the New York Convention.

The numbers chart arbitration's ascent. In 2024 the International Chamber of Commerce (ICC) registered 841 new arbitrations, and the total value of its pending caseload reached a record of roughly US$354 billion, with an average new dispute exceeding US$130 million. The Hong Kong International Arbitration Centre (HKIAC) reported its highest-ever caseload, 352 new arbitrations, up about 25% over the prior year. The London Court of International Arbitration (LCIA), the Singapore International Arbitration Centre (SIAC), and other institutions report similarly robust volumes, the overwhelming majority of them international. Domestically, the American Arbitration Association (AAA) handles more than 100,000 matters a year across its commercial, construction, consumer, and employment dockets. These figures reflect a global consensus that, for many categories of dispute, arbitration is the superior forum.

The Advantages That Drive Adoption

Arbitration's persistence rests on structural advantages that litigation cannot easily replicate—each with limits worth understanding.

Party autonomy comes first and is the most fundamental. In arbitration the parties choose the decision-maker, the procedural rules, the language, the location, and often the substantive law. That control is simply unavailable in national courts, where the judge is assigned, the rules are fixed, and the calendar belongs to the institution. Neutrality follows from autonomy: neither party must litigate in the other's home courts. Helvex and Pacate, headquartered on different continents, can select neutral Singapore as the seat and appoint arbitrators who owe allegiance to neither side.

Expertise is an underappreciated advantage. National judges are generalists who may move from a criminal docket to a custody dispute to a complex commercial case in a single month. Arbitrators can be chosen precisely for their command of the subject matter—a patent dispute heard by a former patent litigator, a construction-delay claim by an engineer-turned-lawyer, a software dispute involving open-source licensing by someone who can actually read source code. In the Helvex matter, a tribunal that understands turbine thermodynamics will grasp the performance evidence far more readily than a generalist judge.

Confidentiality matters to businesses that cannot afford to air sensitive commercial information in open court. Its precise scope varies: English law recognizes an implied duty of confidentiality in arbitration, while many U.S. courts and some other jurisdictions do not imply one, so the clause should address confidentiality expressly where it matters. (A common trap: parties assume arbitration is automatically confidential, then discover that a motion to confirm or vacate the award in court can drag the whole proceeding into the public record.) Still, the private character of arbitration contrasts sharply with the public dockets of most courts—an essential feature for parties protecting trade secrets in remote and cloud environments or proprietary technology.

Enforceability is arbitration's single most powerful advantage in international commerce. Under the New York Convention (1958), ratified by more than 170 states, an award rendered in one member state is enforceable in every other, subject only to narrow grounds for refusal. No comparable regime exists for foreign court judgments; the Hague Convention on Choice of Court Agreements (2005) fills part of the gap but enjoys far narrower adoption. For Helvex, this is decisive: an arbitral award can be enforced against Pacate's assets wherever they sit, in a way a Swiss court judgment could not.

Speed and efficiency are often claimed but require qualification. Arbitration is not inherently faster—complex international cases can run two to four years from commencement to award—but it avoids many bottlenecks of national courts: crowded dockets, mandatory waiting periods, and multiple layers of appeal. Expedited procedures can resolve lower-value disputes in months. Finality, the last advantage, cuts both ways. Awards are final and binding, with very limited grounds to challenge or appeal. For the winner, that means certainty; for the loser, it means living with an outcome that might have been reversed on appeal in court. Parties must weigh that trade-off deliberately before committing—a calculation that belongs in the same column as the case-value and risk-tolerance analysis we set out in Evaluating and Assessing a Civil Lawsuit.

A candid word on cost: arbitration is not cheap. Parties pay the arbitrators and the administering institution in addition to their own counsel, and in high-value cases the tribunal's fees alone can be substantial. The right comparison is not "arbitration versus a free court" but "arbitration's blend of neutrality, expertise, privacy, and global enforceability versus litigation's public, appealable, sovereign-bound process." For the dispute that needs those features, arbitration earns its cost; for a simple domestic collection action, it may not.

The Legal Architecture: Two Regimes, One Logic

Before drafting a single clause, it is worth seeing the legal scaffolding that holds arbitration up. Two regimes do most of the work, and they share a logic: courts should enforce what the parties agreed to and otherwise stay out of the way.

The Federal Arbitration Act in the United States

In the United States, the foundational statute is the Federal Arbitration Act, 9 U.S.C. §§ 1–16 (domestic, Chapter 1), §§ 201–208 (implementing the New York Convention, Chapter 2), and §§ 301–307 (implementing the Panama Convention, Chapter 3). Enacted in 1925 to overcome the old common-law hostility to arbitration, its operative core is § 2, which declares that a written arbitration provision in a "contract evidencing a transaction involving commerce shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract." That last clause—the savings clause—is the hinge of nearly all modern arbitration litigation: arbitration agreements stand on equal footing with other contracts and may be invalidated by generally applicable contract defenses such as fraud, duress, or unconscionability, but not by defenses that single out arbitration for disfavor.

The FAA's machinery follows from § 2. Section 3 requires a court to stay litigation of any issue referable to arbitration; section 4 lets a party compel arbitration where another refuses; section 7 empowers arbitrators to summon witnesses; section 9 directs courts to confirm an award into an enforceable judgment; and section 10 sets out the narrow grounds to vacate one. The Supreme Court has read these provisions to embody, in its much-quoted phrase, a "liberal federal policy favoring arbitration," Moses H. Cone Memorial Hospital v. Mercury Construction Corp., 460 U.S. 1, 24 (1983).

Two structural points trip up the unwary. First, the FAA is substantive federal law that applies in state as well as federal court and preempts conflicting state law, Southland Corp. v. Keating, 465 U.S. 1 (1984)—but it does not itself confer federal subject-matter jurisdiction. A party seeking to compel arbitration in federal court must find an independent jurisdictional basis (diversity or a federal question in the underlying dispute), a quirk the Supreme Court reaffirmed in Vaden v. Discover Bank, 556 U.S. 49 (2009). Second, the FAA carves out a narrow class of transportation workers "engaged in foreign or interstate commerce" (§ 1), whom the Act does not reach—an exemption the Court has read to cover not just interstate truckers but, for example, airline cargo loaders, Southwest Airlines Co. v. Saxon, 596 U.S. 450 (2022), and certain last-leg delivery drivers, while declining to limit it to workers in the transportation industry, Bissonnette v. LePage Bakeries Park St., LLC, 601 U.S. 246 (2024). For most commercial parties the § 1 exemption is irrelevant; for anyone arbitrating with drivers, loaders, or gig-economy delivery personnel, it is the first question to ask.

The New York Convention and the Model Law Internationally

Internationally, two instruments dominate. The Convention on the Recognition and Enforcement of Foreign Arbitral Awards—the New York Convention (1958)—obliges its 170-plus contracting states to recognize and enforce foreign arbitral awards subject only to the narrow refusal grounds in Article V, and to refer parties to arbitration when a valid agreement exists (Article II). The UNCITRAL Model Law on International Commercial Arbitration (1985, amended 2006), adopted in some form by more than 90 jurisdictions (including Singapore, the seat in our hypothetical), supplies a harmonized procedural framework: separability and kompetenz-kompetenz (Art. 16), minimal court intervention (Art. 5), court support for interim measures and evidence (Arts. 9, 17, 27), and a closed list of grounds to set aside an award (Art. 34) that deliberately mirror the Convention's enforcement-refusal grounds. The United States is a Convention state but is not a Model Law jurisdiction; its international cases run on Chapter 2 of the FAA. England is neither—its arbitration law is its own statute, discussed below—but its courts are famously pro-arbitration. The practical upshot: the seat you choose determines which of these regimes supervises your case, which is why seat selection is among the most consequential drafting decisions, treated in its own right below.

The Arbitration Agreement: The Foundation of Everything

Why the Clause Matters More Than You Think

Every arbitration begins with an agreement to arbitrate. Without it there is no jurisdiction, no tribunal, and no award. The arbitration agreement is therefore not a mere procedural detail; it is the constitutional document of the entire process, and its quality determines whether the case proceeds smoothly or bogs down in preliminary fights about jurisdiction, scope, and procedure. The cruel irony of arbitration drafting is that the clause is negotiated when relations are warm and no one expects a dispute, yet it governs precisely the moment when relations have collapsed. Practitioners sometimes call the arbitration clause the "midnight clause"—the one negotiated last, after midnight, when everyone is tired and eager to sign. That is exactly backward: it is the clause most likely to be litigated.

The agreement usually takes one of two forms. An arbitration clause sits inside a broader commercial contract and commits the parties to arbitrate disputes that may arise in the future. A submission agreement (sometimes called a compromis) is a standalone contract entered after a dispute has already arisen, in which the parties agree to submit that specific dispute. Clauses are far more common, because once a dispute erupts and positions harden, parties rarely agree on anything—including how to resolve their disagreement. Helvex and Pacate were wise to include a clause at signing; had they waited, Pacate would have little incentive to agree to arbitrate at all.

Essential Elements of the Arbitration Agreement

A well-drafted clause is concise but complete, addressing each of the following with precision, because ambiguity in any one element can spawn costly satellite litigation. These same considerations apply when embedding dispute-resolution provisions in technology agreements—see our guides to drafting software license agreements and drafting enforceable NDAs for technology transactions.

The clause should define scope—which disputes go to arbitration. Broad language is generally preferred; the standard institutional formulation, capturing "any dispute arising out of or in connection with this contract, including any question regarding its existence, validity, or termination," has been upheld worldwide as reaching virtually all related claims. Narrow language—limiting arbitration to "disputes concerning the interpretation of this agreement"—risks carving out tort, statutory, and validity claims, forcing them into court and creating the nightmare of parallel proceedings in two forums. The clause should fix the number of arbitrators: most rules default to a sole arbitrator for smaller claims and a three-member panel for larger ones, and the choice should track the expected value and complexity of disputes (a sole arbitrator is faster and cheaper; a panel offers broader perspective and a check against idiosyncrasy). It should specify the method of appointment—commonly, in a three-arbitrator case, each party nominates one co-arbitrator and the two nominees select the chair, with the institution stepping in if that fails.

The clause must designate the seat of arbitration—and this is among the most consequential choices a drafter makes. The seat (also called the juridical place) determines the procedural law governing the arbitration (the lex arbitri), which courts have supervisory jurisdiction, and the grounds on which an award can be challenged. It need not be where hearings physically occur; a London-seated case can hold hearings in Dubai for convenience without changing the supervisory law. Popular seats—London, Paris, Singapore, Hong Kong, New York, Geneva—are chosen for pro-arbitration legal frameworks, experienced judiciaries, and developed case law. In the United States, the choice of a domestic seat brings the FAA and the relevant circuit's case law into play, and a thinner layer of state arbitration law (most states have adopted a version of the Uniform Arbitration Act or the 2000 Revised Uniform Arbitration Act) governs purely intrastate matters the FAA does not reach. For Helvex and Pacate, choosing Singapore as the seat means Singapore's International Arbitration Act supplies the lex arbitri and Singapore's courts supervise—an attractive, neutral, strongly pro-arbitration choice.

The clause should set the language of the proceedings (governing submissions, hearings, and the award), since silence invites disputes and costly translation when documents are voluminous. And it should address governing law, distinguishing three potentially different laws: the law governing the substance of the dispute (the lex causae), the law governing the arbitration agreement itself, and the law governing the procedure (the lex arbitri, usually set by the seat). The middle one—the law of the arbitration agreement—is easy to overlook and surprisingly consequential. In the Helvex contract, Swiss law governs the substance while Singapore is the seat; absent an express choice, which law governs the arbitration clause? Different legal systems answer differently, and that very gap recently prompted statutory reform in England, discussed below.

Finally, the clause should choose between institutional and ad hoc arbitration. Incorporating the rules of an established institution—ICC, LCIA, SIAC, AAA/ICDR, SCC, HKIAC—supplies a tested procedural framework and administrative support. Ad hoc arbitration, often conducted under the UNCITRAL Arbitration Rules, dispenses with the institution (and its fees) in favor of party-managed flexibility, but it demands far more active case management and a clause that anticipates breakdowns. For most commercial parties, the institutional route is the safer default.

The Doctrine of Separability

One of arbitration's most elegant principles is separability: the arbitration clause is treated as an agreement separate and independent from the contract in which it sits. The consequence is profound—even if the underlying contract is void, voidable, or never validly formed, the arbitration clause can survive and ground the tribunal's jurisdiction. This prevents a party from escaping arbitration merely by attacking the main contract.

Separability is exactly what stands between Helvex and a procedural quagmire. Pacate's defense is that the whole contract was induced by Helvex's misrepresentations about turbine efficiency and is therefore void. Without separability, Pacate could argue that a void contract takes its arbitration clause down with it, leaving no tribunal to decide anything. With separability, the clause survives the attack on the contract, and the tribunal can proceed to decide whether the contract was in fact procured by misrepresentation. The U.S. Supreme Court adopted this logic in Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395 (1967), holding that a claim of fraud in the inducement of the contract (as opposed to fraud aimed specifically at the arbitration clause) is for the arbitrators. The Court reaffirmed and sharpened the rule decades later in Buckeye Check Cashing, Inc. v. Cardegna, 546 U.S. 440 (2006), holding that a challenge to the validity of the contract as a whole—there, that the contract was illegal and void under usury law—goes to the arbitrator, while a challenge specifically to the arbitration clause goes to the court; and the rule applies even where a party claims the contract was never concluded at all, save where the very existence of the parties' agreement is in question, Granite Rock Co. v. International Brotherhood of Teamsters, 561 U.S. 287 (2010). The English courts reached the same destination in Fiona Trust & Holding Corp. v. Privalov [2007] UKHL 40, reasoning that rational businesspeople intend disputes about their contract's validity to be resolved by the very tribunal they chose for that purpose.

Who Decides Arbitrability: The Delegation Question

Separability answers what happens to the clause when the contract is attacked; a related and equally important question is who decides whether a given dispute is arbitrable at all—the court or the arbitrators. The default answer in the United States is that gateway questions of arbitrability are for the court, because a party should not be forced into arbitration of the very question whether it agreed to arbitrate. But the Supreme Court held in First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938 (1995), that the parties may delegate arbitrability to the arbitrators—if they do so by "clear and unmistakable" evidence. Many courts treat the incorporation of institutional rules (which typically empower the tribunal to rule on its own jurisdiction) as exactly that clear-and-unmistakable delegation. See, e.g., Contec Corp. v. Remote Solution Co., 398 F.3d 205 (2d Cir. 2005); Oracle America, Inc. v. Myriad Group A.G., 724 F.3d 1069 (9th Cir. 2013).

The delegation doctrine has teeth. In Rent-A-Center, West, Inc. v. Jackson, 561 U.S. 63 (2010), the Court applied separability within the delegation clause itself: where the parties delegate arbitrability to the arbitrator, a challenge to the enforceability of the delegation provision goes to the court, but a challenge to the contract or even to the broader arbitration agreement goes to the arbitrator. And in Henry Schein, Inc. v. Archer & White Sales, Inc., 586 U.S. 63 (2019), a unanimous Court rejected the "wholly groundless" exception that some circuits had used to claw back arbitrability; if the parties clearly delegated the question, the court must send even a seemingly frivolous arbitrability argument to the arbitrator. The lesson for drafters is concrete: whether you want arbitrability decided by a court or by the tribunal, say so expressly. Silence leaves the question to a fact-intensive inquiry and to whichever institutional rules you incorporated.

This is the U.S. analogue of the civil-law principle of kompetenz-kompetenz (or compétence-compétence)—the tribunal's power to rule on its own jurisdiction—recognized in Article 16 of the UNCITRAL Model Law, in every major institution's rules, and in the arbitration statutes of most developed jurisdictions. In the Helvex matter, Pacate's contention that the contract is void does not stop the arbitration in its tracks; the tribunal, applying separability and kompetenz-kompetenz, decides for itself whether it has jurisdiction and whether the contract was procured by misrepresentation, subject to the courts of the seat having the final word on jurisdiction at the set-aside stage.

Pathological Clauses: When Drafting Goes Wrong

Not all clauses are created equal, and a surprising number are created badly. A "pathological clause" contains defects that render it ambiguous, contradictory, or unworkable. Familiar pathologies include clauses that make arbitration optional rather than mandatory ("disputes may be submitted to arbitration"); clauses naming a nonexistent institution (referring to "the International Arbitration Association," which does not exist); clauses that demand both arbitration and litigation; clauses prescribing an unworkable appointment mechanism; and the "split clause" that sends some disputes to arbitration and others to a named court without a clean dividing line, inviting a fight about which bucket a claim falls into. Each invites satellite litigation about the dispute-resolution process itself, delaying the real controversy by months or years.

Courts generally try to honor the parties' evident intent to arbitrate, construing ambiguous clauses in favor of arbitrability where they can. See Moses H. Cone, 460 U.S. at 24–25. But that pro-arbitration tie-breaker presumes a valid agreement to arbitrate in the first place; it does not manufacture consent where none exists, and the Supreme Court has cautioned against using it to resolve genuine questions about whether the parties agreed, as opposed to what they agreed to, Granite Rock, 561 U.S. at 301–02. The only reliable cure is prevention: start from a reputable institution's tested model clause, and modify it only where necessary and only with experienced counsel. A clause is cheap to get right at signing and ruinously expensive to fix in a dispute.

Consumer and Employment Arbitration: The Class-Waiver Wars

Most of this guide concerns negotiated commercial arbitration between sophisticated parties. But arbitration's largest footprint in American life is in the adhesive setting—the take-it-or-leave-it arbitration clauses embedded in consumer terms of service and employment agreements—and that setting has generated the most contested arbitration jurisprudence of the past two decades. Any comprehensive account must address it, both because the doctrine is important in its own right and because the same FAA principles that govern Helvex and Pacate govern the consumer who clicks "I agree."

The defining battleground has been the class-action waiver: a clause requiring that disputes be arbitrated individually, foreclosing class or collective proceedings. Plaintiffs argued such waivers were unconscionable or that they impermissibly stripped litigants of the practical ability to vindicate statutory rights. The Supreme Court, in a series of closely divided decisions, came down decisively on the side of enforcement.

The arc begins with Stolt-Nielsen S.A. v. AnimalFeeds International Corp., 559 U.S. 662 (2010), which held that a party may not be compelled to class arbitration unless there is a contractual basis to conclude it agreed to it—silence is not consent. The following year, AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011), held that the FAA preempts a California rule (the Discover Bank rule) that had condemned most consumer class-arbitration waivers as unconscionable; because that rule "stood as an obstacle" to the FAA's objectives and interfered with the fundamental attributes of bilateral arbitration, § 2's savings clause could not save it. Two years on, American Express Co. v. Italian Colors Restaurant, 570 U.S. 228 (2013), closed the most promising remaining escape hatch: even where the cost of individually arbitrating a federal antitrust claim would exceed any individual plaintiff's potential recovery—so that, as a practical matter, no rational plaintiff would bring the claim—the class waiver was still enforceable. The Court rejected the "effective vindication" theory as applied to merely uneconomical (as opposed to legally foreclosed) claims, reasoning that the FAA does not guarantee an affordable procedural path to every remedy, only a forum. DIRECTV, Inc. v. Imburgia, 577 U.S. 47 (2015), reinforced Concepcion by striking down a strained state-court reading of an arbitration clause that had been deployed to evade it.

The capstone, for employment, is Epic Systems Corp. v. Lewis, 584 U.S. 497 (2018). Employees argued that class-waiver provisions in their arbitration agreements violated their right to engage in "concerted activities" under § 7 of the National Labor Relations Act, so that the FAA's savings clause (illegality) rendered the waivers unenforceable. The Court, 5–4, disagreed: the NLRA does not displace the FAA, the agreements must be enforced as written, and individualized arbitration agreements with class waivers are lawful in the employment setting. Epic Systems effectively settled that employers may compel individual arbitration of wage-and-hour and similar claims.

The story has a coda in California, where the Private Attorneys General Act (PAGA) allows employees to sue as private attorneys general for labor-code penalties. In Viking River Cruises, Inc. v. Moriana, 596 U.S. 639 (2022), the Court held that the FAA preempts a California rule barring the division of PAGA actions into individual and representative components: an employer may compel arbitration of the individual PAGA claim, after which—on the Court's reading of state law—the representative claims would have to be dismissed for lack of standing. The California Supreme Court promptly rejected that standing reading as a matter of state law in Adolph v. Uber Technologies, Inc. (2023), preserving the representative claim in state court—a vivid reminder that even after a U.S. Supreme Court arbitration ruling, the state-law contours of who has standing and what survives can shift the practical result. Practitioners tracking this fast-moving area should consult a current case tracker rather than rely on any snapshot; the law of class and representative waivers "is uncertain and rapidly evolving," in the words of one leading practice resource.

Two countervailing developments deserve mention because they limit the reach of forced arbitration. First, Congress acted in 2022 with the Ending Forced Arbitration of Sexual Assault and Sexual Harassment Act (the "EFAA," codified at 9 U.S.C. §§ 401–402), which gives a person alleging sexual-assault or sexual-harassment conduct the option to void a pre-dispute arbitration agreement (and class-waiver) as to that dispute, at the claimant's election. A companion statute, the Speak Out Act (2022), limits the enforceability of pre-dispute non-disclosure and non-disparagement clauses covering such claims. These are narrow carve-outs from the broad pro-enforcement regime, but they are the first significant federal retreat from it in a generation. Second, the Supreme Court has trimmed some procedural advantages defendants had claimed: in Morgan v. Sundance, Inc., 596 U.S. 411 (2022), a unanimous Court held that a party may waive its right to arbitrate by litigating inconsistently, and that no showing of prejudice to the opposing party is required—arbitration agreements get no bespoke, pro-arbitration waiver rule. The throughline of all these cases is the FAA's "equal-footing" command: arbitration agreements are neither specially favored nor specially disfavored; they are ordinary contracts, enforced (and occasionally limited) by ordinary contract principles and the statutes Congress writes.

Arbitral Institutions and Their Rules

Choosing the Right Institution

The choice of institution shapes the procedural character, cost, and duration of the whole proceeding. While parties can arbitrate ad hoc, most international commercial arbitrations are administered by one of a handful of leading institutions, each with its own rules, fee structure, and administrative culture. The table below summarizes the major institutions; because institutions count cases differently and volumes vary year to year, the figures reflect recent (2023–2024) reporting and should be checked against each institution's latest annual report.

The ICC is the most widely used institution for international commercial arbitration. Its signature feature is scrutiny of draft awards by the ICC Court—an internal review that checks each award for formal and substantive defects before it issues, reducing the risk of later challenge and bolstering enforceability. The ICC also requires Terms of Reference, a document defining the issues, the parties' claims, and the procedural framework. Critics say these features add cost and time; proponents say they improve the quality and durability of awards. The LCIA is known for a lighter administrative touch and for charging arbitrators by the hour rather than as a percentage of the amount in dispute (the ad valorem model the ICC uses), which can save money on very large cases but introduces some unpredictability. SIAC has become the leading Asia-Pacific institution, supported by Singapore's pro-arbitration framework and well-regarded judiciary; its rules include an early-dismissal mechanism—akin to summary judgment—allowing a tribunal to dispose of claims or defenses that are manifestly without legal merit or manifestly outside jurisdiction. The AAA and its international arm, the ICDR, dominate the U.S. market; the AAA's Commercial Rules provide tracks for expedited, standard, and large/complex disputes, each with tailored timelines and costs, and its separate Consumer and Employment rules incorporate due-process protocols designed to keep adhesive arbitration agreements enforceable—all of which we examine in our guide to AAA commercial arbitration. JAMS and the International Institute for Conflict Prevention and Resolution (CPR) round out the major U.S. providers, the latter known for non-administered (ad hoc-style) arbitration with institutional rules.

Features Common to Institutional Rules

Despite their differences, the major rule sets share an architecture: all provide for appointing and challenging arbitrators, empower the tribunal to order interim measures (and usually provide an emergency-arbitrator mechanism), set procedural timetables and default conduct rules, and frame the rendering, correction, and interpretation of awards. Several features deserve particular attention.

Emergency arbitrator mechanisms, now standard, allow a party to obtain interim relief before the full tribunal exists. This can be decisive where assets are being dissipated, evidence destroyed, or confidentiality breached—situations that cannot wait weeks for a tribunal to be constituted. If Helvex learns that Pacate is moving to liquidate the assets that would satisfy any award, an emergency arbitrator can order them frozen within days. Expedited procedures compress the timeline for lower-value disputes; under SIAC's expedited track, for example, the award is due within months of constitution, and under the AAA's expedited procedures the hearing is typically limited and the award issued quickly. Multi-party and multi-contract provisions address the increasingly common case involving more than two parties or several related agreements; the ICC, SIAC, and HKIAC rules all provide mechanisms for consolidation of related arbitrations and joinder of additional parties, though the mechanics differ. Institutions also revise their rules regularly—HKIAC updated its rules in 2024 and SIAC's new rules took effect in 2025—so practitioners should always work from the current version.

The Arbitral Tribunal: Selection, Independence, and Powers

Why Arbitrator Selection Is the Most Important Decision You Will Make

If the arbitration agreement is the constitution of the process, the tribunal is its government. The arbitrators decide what evidence to admit, how the hearing runs, and ultimately who wins. In a system with very limited appeal, their judgment is, for practical purposes, final. That is why experienced practitioners devote more attention to selecting arbitrators than to almost anything else. The right arbitrator can streamline a complex case, bring genuine expertise to bear, and write an award that withstands challenge; the wrong one can prolong the proceeding, misunderstand the technology, and produce an award that satisfies no one. For Helvex, an arbitrator who has handled energy-infrastructure disputes and can read a turbine performance curve is worth more than any rhetorical flourish in a brief.

Appointment Methods

Three methods predominate. Party nomination is most common in three-arbitrator cases: each side selects a co-arbitrator and the two co-arbitrators jointly select the chair (who often casts the deciding vote), with the institution appointing the chair if they cannot agree. List-and-strike procedures are typical for sole arbitrators and are the default under the AAA Commercial Rules: the institution circulates a list of candidates, each party strikes the unacceptable names and ranks the rest, and the institution appoints the highest mutually acceptable candidate. Direct institutional appointment occurs when the parties never agreed on a method, the agreed method fails, or the rules so provide; institutions exercise this power cautiously, usually consulting the parties first. A recurring point of friction in party-nominated tribunals is the propriety of ex parte interviews of prospective party-nominees: most rules and guidelines permit a limited interview to confirm availability and absence of conflicts, but not to discuss the merits, and crossing that line is a frequent ground for later challenge.

Independence, Impartiality, and Disclosure

Independence and impartiality are not aspirations but enforceable requirements under virtually every arbitration law and rule set. Independence concerns the absence of financial, professional, or personal ties between the arbitrator and the parties or counsel; impartiality concerns the absence of bias toward a side. Arbitrators bear a continuing duty to disclose anything that might give rise to justifiable doubts about either. The most widely used framework is the International Bar Association's Guidelines on Conflicts of Interest in International Arbitration (revised 2024), which sorts circumstances into a Red List (serious conflicts, the "non-waivable" subset of which cannot be cured even by consent), an Orange List (situations that must be disclosed but do not automatically disqualify), and a Green List (situations needing no disclosure).

A challenge—a formal objection to an arbitrator's continued service—is typically decided by the administering institution or, in some cases, the courts of the seat. The usual threshold is whether the circumstances would cause a reasonable, informed third party to harbor justifiable doubts about impartiality. See UNCITRAL Model Law, Art. 12(2). U.S. courts apply a parallel standard at the enforcement stage under the FAA's "evident partiality" ground, § 10(a)(2); the leading authority, Commonwealth Coatings Corp. v. Continental Casualty Co., 393 U.S. 145 (1968), set aside an award where a neutral arbitrator failed to disclose a significant ongoing business relationship with a party. The disclosure duty has recently been strengthened by statute in England: the UK Arbitration Act 2025, in force since 1 August 2025, codifies an arbitrator's duty to disclose relevant circumstances that might reasonably give rise to justifiable doubts about impartiality—putting on a statutory footing what the common law had recognized in Halliburton Co. v. Chubb Bermuda Insurance Ltd. [2020] UKSC 48 and the IBA Guidelines had long urged.

The Arbitral Process: From Notice to Award

Commencing the Arbitration

The process begins when the claimant files a notice of arbitration—called a Request for Arbitration or Demand for Arbitration depending on the institution—with the administering body and serves it on the respondent. The filing typically identifies the parties, references the arbitration agreement, summarizes the claims and the relief sought, and proposes the number and identity of arbitrators. Under the ICC Rules, the Request must also describe the circumstances giving rise to the claim and indicate the amount in dispute (Article 4). The respondent's Answer—due within roughly 30 days under the ICC Rules, or 14 days under certain AAA rules—should address the claims, raise any jurisdictional objections, and assert counterclaims. As our AAA arbitration guide explains, failing to respond, while not a "default" in the litigation sense, forfeits the chance to frame the dispute and raise defenses early. When Helvex files its Demand seeking the unpaid US$18 million, Pacate's Answer will assert its misrepresentation defense and counterclaim for the cost of the allegedly underperforming turbines.

The Preliminary Conference and Procedural Order

Once the tribunal is constituted, the first substantive step is usually a preliminary conference (also called a case-management conference). Held in person or by videoconference, it is the tribunal's chance to organize the proceeding and the parties' chance to advocate for the procedural framework that best serves them. The conference typically yields a procedural order (or Terms of Reference) addressing the timetable for submissions and hearing; the scope and method of document production; whether written witness statements will stand in for live direct testimony; the treatment of expert evidence; confidentiality; and any bifurcation—splitting jurisdiction, liability, and quantum into separate phases. A party that arrives with a well-prepared draft procedural order usually sets the agenda. This is no formality: the framework established here governs the rest of the case and can decisively influence cost, duration, and outcome. If Helvex wants a fast track to an award, it should press at this conference for tight deadlines and limited document production; if Pacate wants to develop its misrepresentation theory fully, it will push the other way.

Written Submissions and Document Production

The substantive phase opens with detailed written submissions. The claimant files a Statement of Claim—its factual narrative, legal argument, and requested relief, with supporting documents and witness statements. The respondent files a Statement of Defense and any counterclaim. Depending on complexity, a further round (Reply and Rejoinder) may follow.

Document production in arbitration is far narrower than American litigation discovery under Rules 26–37 of the Federal Rules of Civil Procedure. Most international arbitrations follow the IBA Rules on the Taking of Evidence in International Arbitration (2020), which permit requests for specific documents or narrow, well-defined categories that are relevant and material to the outcome. Broad "fishing expeditions" are routinely rejected; the so-called Prague Rules (2018) push even further toward an inquisitorial, document-light model. For litigators steeped in the expansive discovery culture of U.S. courts, this is both a relief and a discipline: you must know what you are looking for before you ask. The contrast is stark enough that the choice between arbitration and litigation is, in large part, a choice about how much discovery you will get—a theme we develop in our [practical discovery refresher](/documents/a_practical_discovery_refresher---mastering_the_tools_rules_ and_pitfalls_of_federal_civil_litigation). One wrinkle worth flagging: U.S. parties to foreign-seated arbitrations once tried to use 28 U.S.C. § 1782 to obtain sweeping U.S.-style discovery in aid of the arbitration, but the Supreme Court closed that door in ZF Automotive US, Inc. v. Luxshare, Ltd., 596 U.S. 619 (2022), holding that § 1782 reaches only "foreign or international governmental tribunals," not private commercial arbitral panels. Where digital evidence is central—as in disputes touching data scraping or platform liability for user-generated infringement—the parties should address electronic-evidence protocols at the preliminary conference to avoid disputes later.

The Evidentiary Hearing

The hearing is the centerpiece. Its format blends common-law and civil-law traditions. Written witness statements, exchanged in advance, usually serve as direct testimony, so the hearing focuses on cross-examination—the common law's great contribution to fact-finding—followed by questions from the tribunal, a feature more familiar in civil-law systems. This saves time by eliminating live direct examination while preserving adversarial testing. Expert evidence follows the same rhythm: reports are exchanged, then experts testify for cross-examination and tribunal questions. In technical cases, tribunals increasingly use "hot-tubbing" (witness conferencing), in which opposing experts testify together, topic by topic, so the tribunal can see precisely where they agree and diverge. In the Helvex hearing, the competing turbine-efficiency experts might be hot-tubbed, letting the tribunal pinpoint whether the shortfall stems from Helvex's design or Pacate's operation.

Arbitration's evidentiary rules are flexible. Unlike a courtroom, an arbitral tribunal is not bound by formal rules of evidence such as the hearsay rule; it has broad discretion over the admissibility, relevance, materiality, and weight of evidence (IBA Rules, Art. 9). That flexibility is a real advantage in technical disputes—including intellectual-property and technology disputes—where the strongest proof may not fit neatly into a court's formal evidentiary channels. The flip side is that an arbitrator's refusal to hear material evidence is one of the few grounds on which an award can actually be vacated (FAA § 10(a)(3); Model Law Art. 34), so tribunals tread carefully when excluding anything a party calls central.

Interim and Emergency Measures

Tribunals have broad authority to order interim measures—provisional orders to preserve rights, prevent harm, or maintain the status quo pending the final award. Common examples include preserving evidence, freezing assets, requiring security, and prohibiting specified conduct. Most major rules now also provide emergency-arbitrator procedures for relief before the tribunal is constituted. Under the ICC Rules, an emergency arbitrator is appointed within two days of an application and must rule within roughly 15 days (Appendix V). Crucially, applying to a national court for interim relief generally does not waive the right to arbitrate; the FAA, the UNCITRAL Model Law, and most institutional rules expressly preserve the agreement to arbitrate even as a party seeks urgent judicial protection. The UK Arbitration Act 2025 further clarifies and strengthens the English courts' powers to support emergency arbitrators and to obtain evidence from third parties. This matters in fast-moving situations—say, suspected trade-secret misappropriation or brand harm in digital-first commerce—where the need for relief may outrun the tribunal-appointment process.

The Award

The award is the tribunal's final, binding decision. It must be in writing, signed by the arbitrators (or a majority), and, under most rules and national laws, must state the reasons on which it rests. The tribunal's remedial authority is broad—monetary damages, declaratory relief, specific performance, injunctions, interest, and the allocation of costs—bounded only by the scope of the arbitration agreement and any mandatory law. Awards take several forms: a final award resolves all remaining issues; a partial award decides some issues (such as liability) while reserving others (such as quantum); a consent award records a settlement and clothes it with the enforceability of an award; and a default award may issue when a party refuses to participate—though the tribunal may not simply rubber-stamp the appearing party's claims and must still require sufficient supporting evidence. If Helvex prevails, it will want a reasoned final award quantifying the unpaid sum plus interest and costs—an instrument it can take across borders to enforce.

Challenging and Enforcing Arbitral Awards

The Limited Grounds for Challenge

Finality is a defining feature of arbitration—and its sharpest risk for the loser. The grounds to set aside an award are narrow and, in most jurisdictions, drawn from Article 34 of the UNCITRAL Model Law or the closely parallel provisions of the New York Convention: invalidity of the arbitration agreement; denial of due process (a party was not given notice or an opportunity to be heard); the award reaching matters beyond the scope of the submission; improper constitution of the tribunal; non-arbitrability of the subject matter under the law of the seat; and conflict with the public policy of the seat. In the United States, the FAA's vacatur grounds (§ 10) are similar in spirit: the award was procured by corruption or fraud; evident partiality or corruption in the arbitrators; misconduct in refusing to hear material evidence or other prejudicial misbehavior; or the arbitrators exceeded or imperfectly executed their powers.

Two American points are essential. First, the Supreme Court confirmed in Hall Street Associates, L.L.C. v. Mattel, Inc., 552 U.S. 576 (2008), that the statutory vacatur grounds under § 10 are exclusive—parties cannot contract to expand judicial review (for example, to add review for legal error), because the FAA's grounds are not a default the parties may rewrite. (Some parties achieve a similar end through a contractual appellate arbitration layer, in which a second arbitral panel reviews the first award for error before it becomes final—a private appeal that the AAA and JAMS now offer by rule.) Second, the once-popular "manifest disregard of the law" doctrine—under which courts occasionally vacated awards where an arbitrator knew the law and consciously ignored it—is now of uncertain vitality after Hall Street; the circuits are split on whether it survives as a judicial gloss on § 10(a)(4) or perished as a freestanding ground. See, e.g., Stolt-Nielsen, 559 U.S. at 672 n.3 (declining to decide). The safe assumption for a drafter is that errors of law or fact, however egregious, are not grounds to vacate unless they amount to a denial of due process, exceed the tribunal's powers, or violate public policy. That is the price of finality, and parties should understand it before they agree to arbitrate. If Pacate loses, it cannot appeal merely because it thinks the tribunal misread the efficiency evidence; it must find a structural defect in the process itself. The flip side—where a court is asked to confirm a domestic award—is nearly automatic: under FAA § 9 the court "must" confirm unless a § 10 ground is shown.

Enforcement Under the New York Convention

Cross-border enforcement is governed primarily by the Convention on the Recognition and Enforcement of Foreign Arbitral Awards—the New York Convention (1958)—with more than 170 contracting states. It is the most effective enforcement mechanism for any adjudicative decision in the world. To enforce, the award creditor presents the court with the authenticated award (or a certified copy) and the arbitration agreement; the burden then shifts to the party resisting enforcement to establish one of the narrow Article V grounds for refusal, which mirror the grounds for setting aside. Courts in most jurisdictions read those grounds narrowly, honoring the Convention's pro-enforcement bias. As the U.S. Supreme Court observed in Scherk v. Alberto-Culver Co., 417 U.S. 506 (1974), the Convention exists to encourage recognition of international arbitration agreements and to unify the standards by which awards are enforced. In the United States, Chapter 2 of the FAA (§§ 201–208) implements the Convention and supplies federal jurisdiction; where the Convention and the FAA conflict, the Convention controls (§ 208). One nuance worth knowing: an award set aside at the seat may, in some jurisdictions (notably France), still be enforced elsewhere, because the Convention's Article V(1)(e) makes refusal on that ground permissive ("may") rather than mandatory—a divergence that occasionally lets a creditor enforce an award that the seat's courts have annulled.

For Helvex, this is the payoff of having chosen arbitration. With a Singapore-seated award in hand, Helvex can pursue Pacate's assets in any New York Convention state—bank accounts in one country, receivables in another—presenting the award and agreement and forcing Pacate to shoulder the burden of resisting. Where a respondent or its assets sit in a jurisdiction with a more complex enforcement environment, additional planning is required; our guide to serving and pursuing foreign defendants in cross-border disputes addresses some of those practicalities.

The table below summarizes the courts' role at each stage.

The Role of National Courts: Supporting, Not Supplanting

National courts and arbitral tribunals exist in a relationship of complementarity, not rivalry. Courts support arbitration by enforcing agreements to arbitrate, appointing arbitrators when private mechanisms fail, ordering interim measures, compelling witness attendance, and ultimately recognizing and enforcing awards. They supervise it by hearing jurisdictional and set-aside challenges and ensuring basic due process. The governing instinct is minimal intervention. The English Arbitration Act 1996, § 1(c), instructs that the court "should not intervene except as provided" by the Act—a principle the 2025 reforms refined rather than abandoned, leaving the 1996 Act as the backbone of English arbitration law. The UNCITRAL Model Law, Art. 5, likewise provides that no court shall intervene except where the law so provides. The FAA, though less explicit, has been read by the Supreme Court to establish a "liberal federal policy favoring arbitration." Moses H. Cone, 460 U.S. at 24.

One procedural detail with outsized practical importance: in the United States, the FAA now generally requires a stay, not dismissal, of litigation when the court compels arbitration. The Supreme Court resolved a circuit split to that effect in Smith v. Spizzirri, 601 U.S. 472 (2024), holding that where a party requests a stay under § 3, the court lacks discretion to dismiss the suit instead. That matters because a stay (unlike a dismissal) is generally not immediately appealable, keeping a reluctant litigant in arbitration rather than handing it an immediate ticket to the court of appeals—an illustration of how the FAA's design steers parties toward, not away from, the arbitral forum.

In practice, this means courts will generally compel recalcitrant parties to arbitrate, stay litigation in favor of a valid arbitration agreement, and enforce awards with very little substantive review, deferring to the tribunal's findings. It is also why the choice of seat is so important: it is the seat's courts that will supervise the arbitration and hear any set-aside application, so parties choose seats whose judiciaries are experienced and reliably pro-arbitration.

Recent Developments and Emerging Trends

The UK Arbitration Act 2025

The most significant recent legislative development in the common-law arbitration world is the UK Arbitration Act 2025, which received Royal Assent on 24 February 2025 and came into force on 1 August 2025, amending the Arbitration Act 1996 for English-seated arbitrations commenced on or after that date. The reform is best described as evolutionary rather than revolutionary, but several changes matter to practitioners. First and most consequential, a new default rule resolves the long-running uncertainty—highlighted by the UK Supreme Court in Enka Insaat ve Sanayi A.S. v. OOO Insurance Co. Chubb [2020] UKSC 38—over which law governs an arbitration agreement: absent an express choice by the parties, the law of the seat now governs the arbitration agreement, and a choice of law for the main contract does not by itself carry over to the arbitration clause. Had Helvex and Pacate chosen London rather than Singapore as their seat, this rule would now answer the governing-law question that their Swiss-law-but-English-seat contract leaves open.

The 2025 Act also gives tribunals an express power of summary disposal—the authority to dispose of a claim or defense that has no real prospect of success, a welcome tool for weeding out hopeless positions early (and a statutory cousin of SIAC's early-dismissal mechanism). It revises the framework for jurisdictional challenges to awards under section 67 (limiting full rehearings where the tribunal has already ruled on jurisdiction), codifies the arbitrator's duty of disclosure discussed above, strengthens arbitrator immunity around resignation and removal applications, and clarifies the courts' powers to support emergency arbitrators and to obtain evidence from third parties. Together, these changes aim to keep England and Wales at the front rank of arbitral seats.

The Digital Transformation of Arbitration

The pandemic forced arbitration to embrace virtual hearings, electronic filing, and cloud-based document management—and the profession has not turned back. Virtual and hybrid hearings are now standard, cutting cost and travel while broadening access; most procedural orders address the hearing platform, cybersecurity, and contingencies for technical failure. Beyond videoconferencing, artificial intelligence is beginning to reshape practice, from document review and legal research to award analytics—capabilities that raise novel questions about the admissibility of AI-generated evidence and the ethical limits of AI-assisted decision-making, questions that intersect with the broader debates we cover in our analysis of generative-AI copyright litigation. The Silicon Valley Arbitration & Mediation Center's 2024 Guidelines on the Use of AI in Arbitration and the ICC's cybersecurity guidance signal the institutions' first attempts to channel these tools; most rules now require parties and tribunals to address data protection and information security in their procedural orders—essential where the dispute itself involves sensitive proprietary data.

Diversity in Arbitrator Appointment

The arbitration community was long criticized for the homogeneity of its leading arbitrators. That is changing. The Equal Representation in Arbitration Pledge, launched in 2015, has drawn thousands of signatories committed to improving gender diversity, and institutions have measurably increased the appointment of women and practitioners from underrepresented regions—ICC Court appointments, for example, approached parity in recent reporting. Parties increasingly insist on diverse candidate slates, both as a matter of fairness and because a broader pool produces better decision-making.

Third-Party Funding

Third-party funding—an external investor financing a party's costs in exchange for a share of any recovery—has grown rapidly and is now accepted in most major arbitration jurisdictions, in both commercial and investor-state disputes. Disclosure of funding is becoming standard: the ICC requires disclosure of the existence and identity of any third-party funder (Art. 11(7)), and other institutions and several seats have adopted similar requirements (Singapore and Hong Kong having legislated to permit funding subject to disclosure). Funding can level the playing field for a meritorious but under-resourced claimant pursuing, say, a patent or intellectual-property claim that would otherwise be unaffordable—but it also introduces a stakeholder with its own financial incentives, which can influence settlement dynamics and strategy. Parties and counsel should understand the funding terms and the disclosure obligations of their chosen seat and institution.

Mass Arbitration

A development unique to the U.S. adhesive setting deserves a flag, even though we treat it more fully in the forum-choice guide. Having insisted on individual arbitration and class-action waivers to avoid class litigation, some companies discovered the strategy's ironic vulnerability: plaintiffs' firms can file tens of thousands of identical individual arbitration demands at once, triggering enormous per-case filing fees that the company itself agreed to pay. This "mass arbitration" gambit has forced institutions (AAA and others) to roll out special mass-filing protocols and fee schedules, and has sent some drafters scrambling to add bellwether and batching procedures to their clauses. It is a reminder that arbitration provisions, like any contract term, can be turned against the party that drafted them.

Expanding Arbitrability

The categories of dispute that may be arbitrated keep expanding. Antitrust claims, once thought non-arbitrable, have been arbitrable in the United States since Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614 (1985); securities claims (Shearson/American Express Inc. v. McMahon, 482 U.S. 220 (1987)), RICO, employment, and consumer disputes followed. More recently, disputes over digital assets and virtual goods, AI-generated inventions, and standard-essential patents and FRAND licensing increasingly head to arbitration, drawn by the need for technical expertise and global enforceability that this guide has emphasized throughout. The frontier of non-arbitrability now sits mostly in bankruptcy (where the public interest in centralized administration can override an arbitration clause) and in the EFAA's sexual-misconduct carve-out discussed above.

Practical Checklist: Preparing for Arbitration

A short, stage-by-stage checklist distills the guidance above.

When drafting the clause: start from a reputable institution's model clause; define scope broadly; specify the seat, language, number of arbitrators, and governing law (including, expressly, the law of the arbitration agreement itself); decide and state expressly who decides arbitrability; consider whether emergency-arbitrator provisions, expedited procedures, or an appellate-arbitration option fit the deal; address confidentiality expressly if it matters; and, in consumer or employment agreements, build in the due-process protections (and fee allocations) that keep an adhesive clause enforceable.

When a dispute arises: read the clause immediately—scope, procedure, time limits, and any preconditions such as mandatory negotiation or mediation; preserve relevant documents and implement a litigation hold; assess whether emergency or interim relief is needed; begin identifying arbitrator candidates with the right expertise; run the same value-and-risk assessment you would for any case (see Evaluating and Assessing a Civil Lawsuit); and prepare a realistic budget.

During the arbitration: invest heavily in the preliminary conference and procedural order; tailor document production to the case, avoiding both over- and under-production; prepare witness statements and expert reports with trial-level rigor; build a hearing strategy around your most impactful evidence; and monitor costs against the budget.

After the award: assess whether any genuine ground for vacatur exists (rarely); if the award is favorable, move promptly to confirm and enforce it in each relevant jurisdiction; and if the award needs interpretation or correction, file the request within the applicable deadline (often 30 days).

Frequently Asked Questions

Is an arbitration clause in a contract really enforceable, even if one side regrets agreeing to it? Almost always, yes. Under FAA § 2, a written arbitration agreement in a contract involving commerce is "valid, irrevocable, and enforceable" and may be invalidated only by generally applicable contract defenses such as fraud, duress, or unconscionability—not by rules that single out arbitration. Buyer's remorse is not a defense.

Can I appeal an arbitration award if the arbitrator got the law wrong? Generally no. Courts review awards only on the narrow statutory grounds (FAA § 10; Model Law Art. 34; New York Convention Art. V)—essentially structural and due-process defects, not legal or factual error. Hall Street confirmed the FAA's grounds are exclusive and cannot be expanded by contract. If you want error review, you must build in a private appellate arbitration layer; you cannot get it from a court.

What is the difference between arbitration and mediation? An arbitrator decides the dispute with a binding award; a mediator only facilitates a settlement the parties themselves agree to and cannot impose a result. The two are often combined in tiered "negotiate, then mediate, then arbitrate" clauses. See our forum-choice guide.

Who decides whether a dispute even belongs in arbitration—the court or the arbitrator? By default, a court decides "gateway" arbitrability. But the parties can delegate that question to the arbitrator by "clear and unmistakable" language (First Options), and incorporating institutional rules often counts as such a delegation. Henry Schein held that a clear delegation must be honored even when the arbitrability argument looks frivolous.

Does suing in court waive my right to arbitrate? It can. Under Morgan v. Sundance (2022), a party that litigates in a manner inconsistent with arbitrating may forfeit the right—and the other side need not show prejudice. Invoke the clause early.

Can my employer or a company force me into individual arbitration and bar a class action? In most circumstances, yes—Concepcion, AmEx v. Italian Colors, and Epic Systems uphold individual-arbitration clauses with class waivers. Important exceptions exist, notably the 2022 EFAA, which lets claimants alleging sexual assault or harassment void such clauses at their option.

How do I enforce an award against assets in another country? Through the New York Convention. Present the authenticated award and arbitration agreement to a court in any of the 170-plus contracting states; the burden then shifts to your opponent to prove one of the narrow Article V refusal grounds. No comparable global regime exists for court judgments—this is arbitration's signature advantage.

Conclusion: Arbitration as a Strategic Choice

Arbitration is not a default. It is a strategic choice, to be made deliberately, with clear eyes about its advantages, its limits, and the procedural landscape in which it operates. For parties who choose well and execute carefully, it offers a combination of party autonomy, expertise, confidentiality, and global enforceability that no single national court system can match. For those who approach it carelessly—with a pathological clause, an uninformed arbitrator selection, or litigation tactics transplanted wholesale into an arbitral setting—it can be slow, expensive, and disappointing.

The difference, as so often in law, is preparation. The best outcomes are built on a well-drafted agreement, a carefully chosen tribunal, thorough case preparation, and a strategic sensibility attuned to the distinctive dynamics of this forum. Helvex's path to enforcing an award against Pacate across borders began not in the hearing room but years earlier, in a single well-drafted clause. Whether you are negotiating that clause in a new contract or managing a multi-jurisdictional dispute before an international tribunal, the same lesson holds: the work you do before the dispute determines how the dispute ends.

Related Articles

• Arbitration, Mediation, and Choosing a Dispute-Resolution Forum: A Practical Guide — how to choose among court, arbitration, and mediation, with deeper treatment of mediation and mass arbitration.
• The Intricacies of AAA Commercial Arbitration — the rules, tracks, and practice of the dominant U.S. institution.
• Evaluating and Assessing a Civil Lawsuit — the case-assessment discipline that applies before you commit any dispute to any forum.
• A Comprehensive Guide to Federal Civil Litigation for Small Businesses — the public-court alternative against which arbitration is measured.
• [A Practical Discovery Refresher](/documents/a_practical_discovery_refresher---mastering_the_tools_rules_ and_pitfalls_of_federal_civil_litigation) — why discovery in court differs so sharply from document production in arbitration.
• Drafting Software License Agreements and Drafting Enforceable NDAs for Technology Transactions — where well-drafted dispute-resolution clauses earn their keep.

This article is for informational purposes only and does not constitute legal advice. Arbitration law and institutional rules differ by jurisdiction and change over time—including, recently, the UK Arbitration Act 2025 and updated institutional rules. For guidance on a specific matter, please consult qualified counsel.

Selected Authorities

Treaties and model instruments: Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York Convention) (1958), Arts. II–V; UNCITRAL Model Law on International Commercial Arbitration (1985, am. 2006), Arts. 5, 8, 11, 12, 16, 17, 34; UNCITRAL Arbitration Rules; IBA Rules on the Taking of Evidence in International Arbitration (2020); IBA Guidelines on Conflicts of Interest in International Arbitration (2024); Hague Convention on Choice of Court Agreements (2005); Uniform Mediation Act (2001).

Statutes: Federal Arbitration Act, 9 U.S.C. §§ 1–16, 201–208, 301–307; Ending Forced Arbitration of Sexual Assault and Sexual Harassment Act, 9 U.S.C. §§ 401–402 (2022); Speak Out Act (2022); Revised Uniform Arbitration Act (2000); Arbitration Act 1996 (England and Wales); Arbitration Act 2025 (England and Wales) (in force 1 August 2025).

Cases: Prima Paint Corp. v. Flood & Conklin Mfg. Co., 388 U.S. 395 (1967); Commonwealth Coatings Corp. v. Continental Casualty Co., 393 U.S. 145 (1968); Scherk v. Alberto-Culver Co., 417 U.S. 506 (1974); Moses H. Cone Memorial Hospital v. Mercury Construction Corp., 460 U.S. 1 (1983); Southland Corp. v. Keating, 465 U.S. 1 (1984); Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614 (1985); Shearson/American Express Inc. v. McMahon, 482 U.S. 220 (1987); First Options of Chicago, Inc. v. Kaplan, 514 U.S. 938 (1995); Buckeye Check Cashing, Inc. v. Cardegna, 546 U.S. 440 (2006); Hall Street Associates, L.L.C. v. Mattel, Inc., 552 U.S. 576 (2008); Vaden v. Discover Bank, 556 U.S. 49 (2009); Stolt-Nielsen S.A. v. AnimalFeeds International Corp., 559 U.S. 662 (2010); Rent-A-Center, West, Inc. v. Jackson, 561 U.S. 63 (2010); Granite Rock Co. v. International Brotherhood of Teamsters, 561 U.S. 287 (2010); AT&T Mobility LLC v. Concepcion, 563 U.S. 333 (2011); Oxford Health Plans LLC v. Sutter, 569 U.S. 564 (2013); American Express Co. v. Italian Colors Restaurant, 570 U.S. 228 (2013); DIRECTV, Inc. v. Imburgia, 577 U.S. 47 (2015); Epic Systems Corp. v. Lewis, 584 U.S. 497 (2018); Henry Schein, Inc. v. Archer & White Sales, Inc., 586 U.S. 63 (2019); Morgan v. Sundance, Inc., 596 U.S. 411 (2022); Southwest Airlines Co. v. Saxon, 596 U.S. 450 (2022); Viking River Cruises, Inc. v. Moriana, 596 U.S. 639 (2022); ZF Automotive US, Inc. v. Luxshare, Ltd., 596 U.S. 619 (2022); Bissonnette v. LePage Bakeries Park St., LLC, 601 U.S. 246 (2024); Smith v. Spizzirri, 601 U.S. 472 (2024); Contec Corp. v. Remote Solution Co., 398 F.3d 205 (2d Cir. 2005); Oracle America, Inc. v. Myriad Group A.G., 724 F.3d 1069 (9th Cir. 2013); Adolph v. Uber Technologies, Inc., 14 Cal. 5th 1104 (2023); Fiona Trust & Holding Corp. v. Privalov [2007] UKHL 40; Enka Insaat ve Sanayi A.S. v. OOO Insurance Co. Chubb [2020] UKSC 38; Halliburton Co. v. Chubb Bermuda Insurance Ltd. [2020] UKSC 48.

Institutional rules and reports: ICC, LCIA, SIAC, AAA/ICDR, JAMS, CPR, SCC, and HKIAC arbitration rules and annual case statistics (2023–2024 reporting). Caseloads vary year to year and institutions count cases differently; consult each institution's most recent annual report for current figures.