Sri Lanka’s conversation about artificial intelligence has quietly crossed a threshold. The question is no longer whether large language models can answer questions. It is whether they can answer them in Sinhala or Tamil, rooted in local law, aligned with school syllabi, and informed by the way Sri Lankan institutions, markets and public services actually function. What is emerging is the idea of a distinctly national model — not a larger brain, but one educated on Sri Lanka.
At the center of this shift is Chat2Find, a technology platform that has spent years assembling the most essential ingredient for such a system: curated, domain-specific knowledge drawn from authoritative local sources. Rather than building a model from the ground up, the platform connects powerful open large language models to live Sri Lankan data systems through retrieval, orchestration and multilingual alignment. The result is not a static intelligence trained once on frozen data, but a dynamic one that continuously consults updated sources.
Within this ecosystem sit several specialized knowledge pillars. LankaLaw houses acts, gazettes, case law, amendments and legal interpretation. LankaTax maps the tax code, circulars, practical examples and compliance guidance. LankaBiz gathers company data, regulatory filings and market intelligence. ExamPazz compiles O/L and A/L past papers with model answers and teacher explanations aligned to the local syllabus. Cheap2Find contributes price intelligence across products, while NewsDive aggregates developments from publishers in real time. Each operates independently as a focused AI search and explanation engine. Together, they form a living, trilingual corpus that a Sri Lanka–focused model requires most: fresh, structured local knowledge.
The innovation is not to merge all this information into a single monolithic model. Instead, Chat2Find is building an orchestration layer in which a master language model determines which expert system to consult for each query. A question about VAT is routed to the tax engine. A query on the 13th Amendment is sent to the legal engine. An economics past-paper explanation draws from the exam system. The model’s role is to retrieve, reason and present the answer in the user’s preferred language and level of understanding.
This architecture keeps knowledge current, as each system updates daily, while enabling responses that feel conversational and locally fluent.
“Chat2Find LLM is right around the corner. When it launches, it will be open weights and open source — free for everyone, forever,” said Avishka Sumanadeera, the 28-year-old founder of Chat2Find. “Sri Lanka’s knowledge infrastructure — our laws, our syllabus, our tax code — should not be locked behind a paywall. We built this for every student, every small business owner, every citizen.”
Beyond its internal systems, Chat2Find has also launched what it describes as the country’s first marketplace for locally built AI tools, AI Mart. The platform allows developers, researchers, teachers, lawyers and accountants to upload their own AI systems trained on Sri Lankan data. These expert AIs become discoverable services within the broader network, expanding the knowledge base beyond a single organization.
As AI Mart grows, the Sri Lanka model will not rely solely on Chat2Find’s databases, but will be able to consult these external expert systems. The design points toward a national AI infrastructure in which local expertise can be integrated and made accessible through a single interface.
A genuine Sri Lankan model must move seamlessly between Sinhala, Tamil and Sri Lankan English. Legal texts, exam material and public information often exist in parallel across languages. By aligning these sources, the system can retrieve content in one language and respond in another, preserving meaning and local terminology. This offers a practical path to trilingual AI without the immense cost of training a foundational model from scratch.
Many of the components are already in place. Domain-specific retrieval systems operate across law, tax, business, education and pricing. Clean Sinhala and Tamil text pipelines have been built using real documents, gazettes and exam materials. APIs allow a coordinating model to query each knowledge base in real time. Early tool-calling workflows enable the system to identify and consult the appropriate Sri Lankan expert engine before generating a response.
The result is an assistant capable of linking a classroom concept to a real Sri Lankan company example, explaining the related tax implications, citing the relevant legal provision and even comparing the current market price of the required textbook — all within a single reply. Such cross-domain reasoning is beyond the reach of any single source or standalone model. It becomes possible only within an integrated ecosystem built deliberately around the knowledge of a nation.
