Inspiration

The Sahyadri mountain range, with its majestic forts, lush trails, and cascading waterfalls, is a magnet for trekkers and nature lovers. But increased tourism has led to overcrowding, safety concerns, and environmental degradation. We were inspired by the Sahyadri Destination Management Organization (DMO), which brings together local guides, conservators, and tour operators to protect this heritage. We wanted to amplify their efforts using AI — empowering both trekkers and stewards with real-time, reliable, and localized information.

What it does

RAG to Ridge is an AI-powered assistant that provides trekkers with real-time, trail-specific guidance. It answers personalized queries — like trail difficulty, seasonal changes, or current trail conditions — with the expertise of local guides. It also supports the DMO with crowd prediction, and knowledge delivery.

How we built it

We grounded our agents using Retrieval-Augmented Generation (RAG), pulling from knowledge articles, live updates, and structured data. We integrated Salesforce Data Cloud to model and predict overcrowding using historical and social media data. We also defined custom objects for experiences, trails, bookings, and stakeholders to enable smooth interaction and reporting.

Challenges we ran into

Data Fragmentation: Consolidating scattered and unstructured trail data

Personalization: Making agent responses context-aware and site-specific

Scalability: Designing for a range of sites with varying complexity and resources

Accomplishments that we're proud of

Created agents that match the expertise of local guides

Developed a predictive model for site overcrowding

Built a collaborative ecosystem between trekkers, guides, and the DMO

Bridged tech and tradition to support heritage and nature

What we learned

We learned that the real power of AI comes from grounding it in local context. Building for people who live and work in these regions helped us shape solutions that are relevant, respectful, and impactful.