← Back to HAQQ Blog

The Middle East Legal Tech Landscape: A 2026 Field Guide

By HAQQ Team · · 16 min read · Mena

Gulf governments are racing ahead on legal AI while the Arabic data underneath stays scarce. A current, data-backed field guide to MENA legal tech — who's building, who's buying, and where the gap really is.

AI Is Already in the Courtroom

In February 2026, a court in Abu Dhabi ordered a group of lawyers to pay 282,508 dirhams - about $77,000 - for citing cases that didn't exist.

The cases were invented by an AI tool, dropped into a defence submission, and filed without anyone checking whether the authorities were real. The court - the ADGM Court of First Instance, which runs on English common law inside the United Arab Emirates - called the conduct 'reckless' and made the former legal representatives eat the wasted costs.

We keep coming back to that case, because it tells you almost everything about where legal tech in the Middle East actually is right now. Not where the conference panels say it is - where it *is*. AI is already in the courtroom here. It's already shaping filings, already producing winners, and already producing people who pay 282,508 dirhams to learn that a language model will lie to you with a straight face.

Why Everyone Misreads This Market

We've spent the last two years building and selling legal AI in this region, and we've watched a lot of smart people - investors, founders, journalists, even lawyers who work here - get the Middle East wrong. They treat it as a smaller, later, slightly dustier version of the US or UK legal-tech market. A place that will eventually catch up.

That framing is just wrong. The Middle East isn't behind. It's **ahead in some places and years behind in others, at the same time**, and the two facts are knotted together in a way we haven't seen in any other market. The governments are sprinting. The data underneath them is a swamp. Both things are true, and most 'landscape' articles only tell you one of them.

So here's the whole map in one place: who's building, who's buying, who's moving in from outside, and why the hardest problem in legal AI on Earth might be the one nobody's funding. (One housekeeping note: we're going to skip the headline market-size numbers you'll see quoted elsewhere - when we tried to trace them, most led back to nothing. We'd rather show you real companies and real data than a confident-looking TAM chart built on air.)

Governments Are the Aggressive Adopters

In most of the world, regulators are the brake. In the Gulf, they're the accelerator. This is the part that genuinely surprises people the first time they see it.

Start with the headline. In April 2025, the UAE Cabinet approved a **Regulatory Intelligence Office** - an AI system that links every federal and local law to court rulings and public services, tracks the real-world impact of legislation, and *recommends updates to the law itself*. The government's own pitch is that AI will speed up the legislative process by up to 70% (a target, not a measured result). Humans keep the final sign-off - but read that back slowly: a national government is building a system to help write and continuously revise its own statutes. The closest analogues anywhere else are academic. Here it's a cabinet decision.

Then there are the courts, where it stops being a press release and starts being infrastructure. Saudi Arabia's Ministry of Justice runs **Najiz**, an e-justice portal with around 160 services, and **Nafith**, an e-enforcement platform whose 'virtual enforcement court' the ministry says collapses enforcement from twelve steps to two. The UAE keeps one-upping itself: a Ministry of Justice 'virtual lawyer,' and a **Court of the Future** unveiled at GITEX 2025 that the government claims will cut litigation time by up to 90%. (We'd bet against 90%. The direction is the point.) Qatar's QICDRC and the DIFC Courts have gone further still - they wrote formal rules for AI use in litigation, in 2026 and 2023 respectively, *before* most US state bars finished arguing about whether ChatGPT counts as the unauthorized practice of law.

**Government AI-in-law in the Gulf, a quick timeline:**

If you only read one section of this guide and walked away thinking 'the Middle East is a fast follower,' you'd have it exactly backwards. At the *government* layer, parts of the Gulf are the global frontier.

The Common-Law Islands

Here's a piece of context that trips up almost every outsider. The Gulf is mostly **civil law** (codified statutes, the continental-European tradition), with Islamic **Sharia** governing personal-status matters like marriage, inheritance and family. But sitting inside that civil-law world are a handful of *common-law islands*: financial free zones with their own laws, their own courts, and English as the language of business.

The DIFC in Dubai and the ADGM in Abu Dhabi both run English-language, common-law courts - ADGM applies the common law of England directly, by statute. Qatar has its own version in the QFC and QICDRC. These zones were built to make international business feel at home, and for legal tech they matter enormously - in a slightly cruel way.

The common-law islands are where the digital-court infrastructure is most advanced, and they're English-language and internationally staffed - which means a Harvey or a Legora trained mostly on US and UK law can actually be useful there on day one. And that's the cruelty: the slice of the regional market where Western legal AI works out of the box is the small, English-speaking, common-law slice. The vast majority - the Arabic-language, civil-law, Sharia-inflected mainland, where most of the actual legal work happens - is the part those tools handle badly. The hallucination case we opened with happened on the *easy* side of that line.

Who's Actually Building

So who's building for the hard side? A small but real and growing group of companies, most of them under-funded relative to the problem, spread across the region. The geography tells a story.

**The UAE** has the most mature commercial layer - Lexzur (practice management and contract tooling, fully Arabic-configured) and Clara (automated company formation across ADGM, DIFC, Cayman and Delaware). **Saudi Arabia** is where the homegrown energy is loudest, and it's almost all Arabic-first: Shwra (an AI assistant answering on Saudi law), Signit (the homegrown contract and e-signature platform, hosted in-Kingdom and wired into national identity rails), and Mohami. Notice these are built *on top of* the government infrastructure from the last section - in-Kingdom hosting and national-ID integration aren't features here, they're table stakes.

**Egypt** is a big legal market with a tiny legal-tech sector - Elmetr (a lawyer marketplace) and Waseya (an Islamic-inheritance and will-making tool, which is a genuinely Sharia-native wedge). **The Levant** punches above its weight: Lexyom and HAQQ out of Beirut, and Qistas out of Amman (Arabic legal research across nine-plus jurisdictions). **North Africa** has the quietly impressive Legal Doctrine, an Algiers startup that became the 'Google of Algerian law' and expanded across 22 African jurisdictions, plus Tunisia's Qanouni.

**Homegrown legal tech across MENA, by country:**

The pattern across all of it: the rounds are small. Most of these companies are at pre-seed or seed. The homegrown layer is real, it's serious, and it is dramatically under-capitalized for the size of the problem it's chewing on.

The Global Players Moving In

The giants have noticed. The single most telling fact in this whole landscape: one law firm anchored *both* of the leading Western legal-AI companies in the region. **Al Tamimi & Company**, the largest full-service firm in the Middle East, became the region's first firm to partner with Harvey - and then, in November 2025, became Legora's first Middle East partner too, with a firmwide rollout and a purpose-built *Arabic interface*. When the same buyer onboards both of your fiercest competitors, you've got a market being courted hard and a customer that knows it has leverage.

LexisNexis has been here for years with Lexis Middle East (bilingual research across the GCC) and has layered its Protégé AI on top. The arms race nobody's narrating clearly is the **Arabic** one: both Harvey and Legora are now investing specifically in Arabic capability for this market. That's the tell. The Western incumbents have figured out you cannot win the Middle East with an English-only product - because the language is where the entire game is decided.

The Money Doesn't Match the Ambition

Put the two halves side by side and the asymmetry is almost comic. Demand is lit up - a 2025 Deloitte survey of senior tax, finance and legal leaders across the GCC found the share using generative AI jumped from roughly **48% to 71% in a single year**. (That figure covers all three functions combined, not legal alone, but the trend line is steep either way.)

**Disclosed MENA legal-tech rounds (selected):**

Harvey and Legora, by contrast, raise at multi-billion-dollar valuations - literally off this chart. The largest disclosed homegrown round we could find is in the low single-digit millions. That gap is the opportunity, not the obituary - but it's real.

The Hard Part: Arabic Legal Data

Now the swamp. This is the section the panels skip, and it's the single most important thing to understand about legal AI in this region: building it in Arabic is **genuinely, structurally harder** than building it in English. Not 'needs a bit of localization' harder. Different-order-of-magnitude harder.

**One: the language fights you.** Arabic is morphologically explosive, usually written without the short-vowel diacritics that disambiguate meaning, and legal text (formal Modern Standard Arabic) differs sharply from the dialects clients actually speak. A tool here has to read formal statutes *and* understand someone describing their problem in Gulf or Egyptian or Levantine dialect.

**Two: the data barely exists in usable form.** In the US, if you want to ground a legal model you reach for CourtListener - over ten million court opinions, free, with a bulk API - or Pile of Law, a 256-gigabyte open legal corpus. The Arabic-language equivalent does not exist. When researchers built the first serious Arabic legal-reasoning benchmarks, they had to scrape a single country's commercial-court cases and say plainly that the field 'remains under-explored.' We're talking thousands of cases, versus the West's tens of millions. A lot of regional law still lives in scanned PDFs of official gazettes - image-only, un-searchable until someone runs OCR.

**Open legal data: the West vs. Arabic.**

**Three: it's three legal systems at once.** A single tool serving, say, the UAE has to straddle Arabic Sharia-grounded reasoning for personal status, Arabic-and-English civil-commercial codes for business, *and* English common-law reasoning in the DIFC and ADGM. Three traditions, two languages, one country. No off-the-shelf model was trained for that combination, because that combination only exists here.

The encouraging counterweight: the *foundation* layer is getting strong fast. The Gulf has poured serious money into sovereign Arabic LLMs - Jais (G42/MBZUAI), Falcon (Abu Dhabi's TII), ALLaM (Saudi's SDAIA), Fanar (Qatar). These are excellent general Arabic models. But none of them is a *legal* model. They're the engine block. Somebody still has to build the legal car - the corpus, the structure, the jurisdiction-specific reasoning - on top. That's a domain problem, not a compute problem. You can't buy your way out of it with a bigger GPU cluster.

HAQQ's Take

We should be straight about where we sit, because we've been quoting this market while standing in it. HAQQ is a legal-AI company founded in Beirut in 2022 by Antoine Kanaan and Maître Abbas Kabalan, building a vertically integrated legal system - drafting, research, review, case management - tuned to produce jurisdiction-specific output rather than generic text. So when we tell you the data-and-language layer is the real battleground, understand it's also the bet the company is making. Read accordingly.

Early on, we ran our own version of that ADGM lawyer's experiment internally: we fed leading general-purpose tools a handful of regional matters and watched them invent provisions, miss Sharia-specific rules, and confidently apply the wrong jurisdiction. It wasn't that the models were bad. It's that they'd never seen the law. That afternoon basically became our product thesis.

Here's what we actually believe. The Middle East will be one of the most important legal-AI markets in the world - not because of oil money looking for a home, but because the *top-down* conditions are almost uniquely favorable: governments that deploy AI in courts faster than any Western democracy, regulators who write the rules early, sovereign investment in Arabic models, and a young, mobile-first professional class. The demand side is lit up. The supply side - working, Arabic-and-Sharia-and-common-law-aware tooling - is where the gap is. Whoever fills that middle, patiently, jurisdiction by jurisdiction, wins a market the governments have already pre-warmed.

The lawyers in that Abu Dhabi courtroom learned the hard way that the tools aren't there yet. The 282,508-dirham lesson wasn't 'AI is bad.' It was 'the generic tool doesn't know your law.' That distinction is the entire opportunity in the Middle East right now - and we think the race gets decided in Arabic, and decided here.

Key Takeaways

Sources & Further Reading