Legal technology software in the Middle East is moving beyond lawyer discovery, with digital intelligence emerging as a new gateway to legal knowledge. HAQQ, an AI legal platform, is placing its heftiest bets on that shift. What was initially a marketplace connecting users with lawyers, later morphed into an AI legal platform built around Arabic and civil law systems.
With more than 15,000 law firms across 80 countries having used the platform, HAQQ is now tackling a bigger challenge of administrating sound legal guidance while navigating the linguistic, legislative, and jurisdictional complexities global legal AI platforms were not originally built to address.
In a conversation with Inside Telecom Magazine, HAQQ Founder and CEO, Antoine Kanaan, walked through what that shift actually required. It wasn’t simply about layering AI into existing directory, but rethinking how legal knowledge is structured, retrieved, and finally, delivered in a region where the dominant global models were trained on legal traditions that bear little resemblance to the ones governing daily life in the Middle East.
Civil law systems, Arabic linguistic complexity, and the jurisdictional fragmentation of a market of a multi-jurisdictional footprint across 22 countries are not problems that a model fine-tuned on Anglo-American case law solves out of the box.
1-Why did you leave the marketplace model behind?
Kanaan actually started as a platform that connected people with lawyers. Then we launched practice management software for law firms, and later we added AI. What we learned was simple: people don’t wake up wanting to find a lawyer, they want an answer to a legal question. Finding a lawyer was not the biggest problem. The real challenge wasn’t finding a lawyer, but getting reliable legal answers quickly, affordably, and from a source people could trust. We also realized that the marketplace model wasn’t a great fit for many Middle East and North Africa ( MENA) markets.
Legal services aren’t something people buy regularly, so marketplaces struggle to build the kind of repeat activity they need. And lawyers weren’t looking to buy more leads. They wanted technology that saved them hours of work. So, instead of focusing on connecting people, we focused on helping produce the legal work itself.
2-What’s actually built in-house?
Everything is built in-house. We designed and developed the entire intelligence layer ourselves, from the way the system understands a legal question to how it identifies the relevant jurisdiction, verifies the sources, and generates reliable answers or legal documents. We don’t build the underlying large language models themselves.
Instead, we integrate the world’s leading AI models, such as OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini, into technology we’ve built from the ground up. Our value isn’t the model, but everything around it. We’ve engineered a legal reasoning system that transforms AI into an assistant that lawyers and legal professionals can trust.
3-How does it handle different legal systems?
We use one core platform that can be adapted to different countries rather than rebuilding everything from scratch each time. Before generating an answer, the system first identifies which country’s laws apply. Every answer is linked back to the actual laws it comes from, so users can verify it themselves. Because many countries in our region rely primarily on written legislation rather than case law, we focus on the legal codes themselves.
If the law isn’t clear or different sources point in different directions, the platform highlights that instead of pretending there is only one answer. And when confidence is low, it recommends human review.
4-What’s your edge in Arabic and civil law, and how do you measure it?
Most legal AI tools were built for English and common law systems such as the US, or UK. Our region works differently. The law is different, the language is different, and the way lawyers work is different. We built HAQQ with Arabic and civil law at its core from day one, not as something added later.
We constantly test the platform across countries like Saudi Arabia, Egypt, Lebanon, Qatar, and the UAE. But we don’t just measure whether it gives an answer. We also measure whether it knows when not to answer. In legal AI, knowing your limits is just as important as being right.
5-Who designed and graded your benchmarks?
We designed and ran the benchmarks ourselves, and we are very open about that. They haven’t yet been independently audited. To make the comparisons fair, every system receives the same prompts under the same settings. Before testing, we lock the scoring criteria, which includes legal accuracy, citation quality, hallucination rates, and whether the legal references actually exist. An AI model performs the initial scoring, and then our legal team reviews the results. We’re happy to publish the methodology and even run live comparisons using prompts chosen by journalists or customers.
6- Who makes up HAQQ’s user base?
Today, more than 15,000 law firms have used HAQQ across over 80 countries, with more than 800 paying customers. The biggest surprise for us was who uses it. We originally built HAQQ almost entirely for lawyers. Today, around 80% of new users are people without legal training who simply need clear answers to legal questions. That confirmed something we had suspected from the beginning where access to legal information is a much bigger problem than access to lawyers.
More than 80% of our users find us organically rather than through advertising. One example is the Government of Louisiana, which discovered the platform on its own and became a customer without us ever approaching them.
7-Where is the money going?
The funding is primarily going into product development, expanding our engineering team, deploying projects we’ve already signed, strengthening security and compliance, and growing in new markets. One advantage of our enterprise and government model is that infrastructure costs are often covered by the customer. In many of those deployments, the partner provides the servers or (Graphics Processing Unit) GPU infrastructure, while we provide the software, AI capabilities, and ongoing support. So, we’re spending our money on building and deploying the product, not buying expensive infrastructure.
Today, we have around $1.5 million in live Software as a Service (SaaS) revenue, about $11 million in signed or ministry-approved projects, and a pipeline worth roughly $235 million. Of course, signed contracts aren’t the same as collected revenue, and we think it’s important to be transparent about that distinction.
8-Which certifications do you actually hold?
Security is one of the first questions enterprise customers ask us. We are compliant with Security Operations Cente (SOC) 2 Type II, ISO 27001, ISO 42001, and General Data Protection Regulation (GDPR). We also never use customer data to train AI models.
9-Information or advice – and who’s responsible?
Our goal isn’t to replace lawyers. It’s to replace uncertainty. Again, we’re a software company, not a law firm. If someone has a legal question, they should start with information they can trust and then involve a lawyer when legal judgment is actually needed.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Interviews to stay informed and up-to-date with our daily articles.
