SIH26045: IP-SAKTI Sahayak a multilingual, RAG-based (source-cited) AI assistant for Intel — SIH 2026 | Wynit

SIH26045: IP-SAKTI Sahayak a multilingual, RAG-based (source-cited) AI assistant for Intel — SIH 2026 | Wynit

What they'll ask you

12

Generated from the brief's own wording and the competition's published rules, never from a guess about what this ministry prefers.

Specific UI/UX wireframes or designs?
The brief never answers this, so a panel will. Whatever you decide, say it the same way twice.
List of exact databases and URLs to scrape for the initial corpus?
The brief never answers this, so a panel will. Whatever you decide, say it the same way twice.
Has any part of this been shown at a previous event, hackathon or college project?
The guidelines are explicit: your solution must not have appeared in any previous event or programme, of any sort. A recycled project is what a team under time pressure reaches for.

9 morethe ones a panel gets to if there is time

The brief asks for formulation classification flow. How would you build that?
Decoded from SIH26045 itself: A questionnaire module that asks minimum clarifying questions to classify an Ayurvedic product into categories like classical medicine or new drug. The brief asks for it by name.
The brief asks for jurisdiction toggle. How would you build that?
Decoded from SIH26045 itself: A UI component that switches between national and international legal frameworks and keeps the answer sets visibly separate. The brief asks for it by name.
The brief asks for rag and source citation engine. How would you build that?
Decoded from SIH26045 itself: A retrieval-augmented generation engine that answers queries while citing the specific statute, treaty article, or registry record. The brief asks for it by name.
The brief asks for multilingual delivery interface. How would you build that?
Decoded from SIH26045 itself: An interface supporting national languages, leveraging language infrastructure such as Bhashini. The brief asks for it by name.
The brief asks for guardrails and escalation. How would you build that?
Decoded from SIH26045 itself: A system providing mandatory legal disclaimers, safe abstention on uncertain queries, and an escalation path to a human IP facilitator. The brief asks for it by name.
Why not use what already exists? Name the closest thing to this that is already running.
A team that has not named the alternative themselves is answering this for the first time in the room.
Which single thing will you demonstrate end to end, start to finish, with nothing skipped?
Ours, not a rule: a narrow thing that fully works survives questioning better than a broad thing that half works. If nobody on the team can name it, that is the finding.
Show me this working: the jury will check answer accuracy, citation correctness, safe abstention on out-of-scope queries, and multilingual quality.
This is the evaluator read for your problem statement, decoded from the brief's own wording.
Show me this working: they will also test if the jurisdiction toggle keeps national and international rules separate.
This is the evaluator read for your problem statement, decoded from the brief's own wording.

The government's original description (685 words)

Background: Ayurveda rests on a vast corpus of codified and community-held traditional knowledge (TK) and on therapeutics derived from plant, microbial and animal sources. Protecting and commercialising an Ayurvedic product means navigating several overlapping regimes at once: patents, geographical indications (GI), trademarks, copyright, designs, trade secrets and plant-variety rights; the Access-and-Benefit-Sharing duties that flow from Indiaâ€™s sovereignty over its biological resources; and the drug-regulatory framework that decides whether a formulation is a classical medicine, a proprietary medicine, a new drug, a phytopharmaceutical, a food or a cosmetic. Practitioners, researchers, AYUSH startups and MSMEs and cultivators routinely struggle with this. The result is twofold: legitimate Ayurvedic innovation is under-protected and under-commercialised, while Indiaâ€™s traditional knowledge remains exposed to misappropriation abroad. Recent shifts â€” the 2024 patent and biodiversity rules, the WIPO Treaty on Genetic Resources and Associated Traditional Knowledge (2024) and a fast-moving advertising and regulatory landscape â€” make authoritative, plain-language guidance more necessary than ever, yet no such tool exists for the AYUSH community. Description: The assistant answers IPR questions specific to Ayurveda with accuracy, source citation and jurisdictional clarity, keeping the national and the international layers distinct through an explicit jurisdiction switch so that answers are never conflated. Because intellectual property for an Ayurvedic product is inseparable from how the product is regulated, the assistant first helps classify the formulation. It asks the minimum clarifying questions to determine whether the product is a classical/generic medicine (formulation and method drawn from a First-Schedule authoritative text), a patent-or-proprietary medicine, a new or non-classical drug requiring proof of safety and effectiveness, a phytopharmaceutical, an Ayurveda-Aahar / nutraceutical, or a cosmetic â€” and then states what each category requires and its very different IP and ABS posture. For example, a classical formulation is largely traditional knowledge that faces the Section 3(p) patenting bar and is defended through the Traditional Knowledge Digital Library, whereas a new drug gains genuine patent potential but must generate clinical evidence. National coverage spans the Patents Act (and the 2024 Rules), the GI, Trade Marks, Designs, Copyright and Plant-Variety regimes, the Biological Diversity Act (as amended in 2023, with the 2024 Rules) and the allied drug, advertising, labelling and food/cosmetic regimes â€” the Drugs and Cosmetics Act, the Drugs and Magic Remedies (Objectionable Advertisements) Act and the FSSAI Ayurveda-Aahar regulations. International coverage separately spans TRIPS, the Convention on Biological Diversity and the Nagoya Protocol, the WIPO GRATK Treaty, the PCT, the Madrid and Hague systems, the Budapest Treaty (for micro-organism deposits) and the herbal-product market-access regimes of key export markets. The assistant also facilitates access to authoritative sources â€” free official databases directly and the userâ€™s own paid subscriptions only with explicit, logged permission â€” so that a user can move from a question to the right registry, record or form. It must cite the specific statute, rule, treaty article or record it relies on; clearly state that it provides information and not legal advice; keep its corpus current as the law changes; and never fabricate authority. Expected solution: A deployable, multilingual assistant built on retrieval-augmented generation grounded in a curated, version-tracked corpus of statutes, rules, treaties, pharmacopoeial standards, registry records and case law, so that every answer is traceable to a source and hallucination is minimised. The solution should provide: a jurisdiction toggle (India vs international) with the two answer-sets kept visibly separate; routing across IP types together with the formulation-classification flow; an ABS-compliance helper and a TKDL / prior-art pointer; mandatory source citations with a confidence indicator and a path to escalate to a human IP facilitator; multilingual delivery (leveraging national language infrastructure such as Bhashini); and guardrails, a standing 'information, not legal advice' disclaimer and privacy, audit and security aligned to the Digital Personal Data Protection regime and to recognised AI-application standards. A relational knowledge graph and agentic, multi-source orchestration deepen multi-step reasoning and the build can be staged â€” a citation-grounded retrieval MVP first, then the graph and agentic layers, then paid-source connectors and the full multilingual and voice experience. The output should be evaluable on answer accuracy, citation correctness, safe abstention on out-of-scope or uncertain queries and multilingual quality.