Lexify

UI of Lexify

Inspiration

This initiative was inspired by a real-world problem that hit close to home. A close friend of ours had just opened a startup, and in India, startups face an overly complex compliance system, all of which he was unaware of. He did have some initial registrations but lacked complete knowledge of the industry filings and other important documents required for his business. When he missed license renewals and statutory documentation, he faced a business lawsuit, and the resulting penalties pushed him to the brink. A few weeks later, while sitting in that same office watching him struggle with stacks of legal documents, we realized the true extent of compliance failure across MSMEs in India. What if there was an AI assistant to handle all legal and regulatory paperwork for businesses automatically? This would prevent penalties and allow entrepreneurs to focus entirely on scaling their ventures.

*This vision, simple yet impactful, gave birth to our AI-powered MSME Legal & Compliance Support Portal: Lexify. *

What it does

We are building a Unified AI-led MSME Growth & Compliance Platform that allowed MSMEs to effectively comply with regulations, leverage publicly available data, and make data-driven decisions on their growth journey. This solution is a combination of Granite LLM models provided by IBM tech assets for intelligent text processing with Watsonx Orchestrate ADK for workflow automations to ensure that MSMEs remain compliant. This whole process was carried out in few phases-

We collected information from government MSME portals, regulatory websites at the state level, FSSAI, GST, industry boards, and market trend APIs.
We utilized Granite models to classify industry regulatory obligations, summarize government notifications in plain language, and determine market demand, implementing policies into actionable insights specifically for each MSME sector.
3.We are using Watsonx Orchestrate ADK to create sector-specific compliance checklists, generate automated reminders for license renewals or tax filings, deadlines as well as suggest government schemes based on their particular case.

How we built it

Python for Data scraping, ML/AI, backend APIs JavaScript (React) for Frontend UI integration IBM watsonx.ai for AI/LLM models (data analysis, insights) IBM watsonx.orchestrate ADK will be used in Workflow & task automation IBM API Connect to Secure API management Flask to Customize API development IBM Cloud for the Hosting & scaling DB (IBM Db2) will Store MSME & processed data

Challenges we ran into

Access to Government Data – Different government agencies such as FSSAI, GST, Labour, and others, have their own separate portals and they do not integrate into one API. In addition, updates often come in PDFs or scanned notices of many pages, making clean data extraction difficult. Frequent Rule Changes – Rules and deadlines keep changing. We needed a system that adapts automatically without hardcoding. Understanding Legal procedures – Government notifications are full of legal documents and updates. Training the AI to simplify this into actionable IF–THEN rules was tough. Standardization of MSMEs Data – MSME data is received in varied formats (Excel, Tally, paper bills, paper scans). Bringing them to AI-compatible format was a major challenge.

Accuracy & Trust – There is always the possibility for small, seemingly inconsequential AI mistakes that can lead to serious compliance breaches and penalties. We are developing strong validation, human-in-the-loop review, and fallback systems to make sure that the system is dependable.

Accomplishments that we're proud of

The legal sectors are still untouched by many and even our little progress in it felt appreciable.

What we learned

We gained knowledge and experience with government compliance data, which is often unstructured, incomplete, and difficult to standardize. 2.We learned the process of converting complex legal and regulatory language into structured rules that can be analyzed by AI systems.
Built the ability in AI system to adapt quickly to changing requirements and frequent regulatory updates, ensuring the system remains flexible and scalable.