We built the citation layer that legal AI was missing
AI legal tools were generating answers without sources. Practitioners in the UAE and MENA were expected to use those answers in advice without a way to verify them. We set out to fix that problem at the source.
Source attribution is not a feature. It is the whole point.
Every statement a lawyer uses in advice has to be traceable to its source. That is not a preference; it is a professional requirement. We built Qanooni so that every answer the system returns comes with a numbered citation to the specific provision, article, and law it is based on.
We focused on UAE and MENA law because the legal landscape here is genuinely complex: federal law, free zone frameworks, dual common-law jurisdictions, and GCC regulatory coordination all intersect on the same transaction. Generic AI tools that were built for US or UK law get the local picture wrong. We built our corpus and our extraction model for this region specifically.
Qanooni was founded in Dubai in 2024 and is angel-backed. We work directly with early-access legal teams whose feedback shapes the product.
See how it works
Three principles we build by
If the system cannot trace its answer to a specific provision in the corpus, it does not return the answer. We built the citation constraint into the architecture rather than treating it as a display layer on top of a generative output.
UAE law, DIFC common law, ADGM frameworks, and GCC regulatory requirements are distinct from any other legal system. We built our corpus for this region and test our citation accuracy against local source texts.
The people using Qanooni are responsible for the advice it helps them produce. Every design decision we make asks: does this help the practitioner exercise their judgment more reliably, or does it shortcut that judgment?
Who builds Qanooni
Anuscha built her career working with legal teams on the software that runs their practice workflows. Over the years, she saw where those tools consistently failed: research outputs arrived without sources, leaving lawyers to manually trace each conclusion back to primary law before they could rely on it. She co-founded Qanooni to put citation-first retrieval at the center of the research workflow rather than treating it as an optional layer added after the fact.
Tariq has spent years building retrieval and document-processing infrastructure for high-stakes professional environments where accuracy per output is not negotiable. At Qanooni, he designed the citation extraction architecture, which ties each statement in a research response to its precise source location in the corpus at inference time rather than attaching citations as a post-generation step.
Priya practiced as a solicitor before moving into product. She spent her time in practice noticing which software decisions made legal work more reliable and which created more work to compensate for them. She joined Qanooni to translate the citation-first research concept into a tool that fits the actual workflow of in-house counsel, not a separate system that competes with it.
Questions or want to discuss a Firm plan?
We work directly with corporate legal teams and law firms to configure Qanooni for their practice. Get in touch to talk through what that looks like for your team.