jameschaearley

I've created a conversational weather assistant that leverages Snowflake Cortex's AI capabilities with Claude 3.5 Sonnet. The application allows users to ask natural questions like "Is it going to rain in Seattle today?" or "Should I wear a jacket in Chicago?" Claude handles calling the right APIs, parsing responses, and providing straight answers without requiring users to decipher weather jargon. Key features:

Leverages Claude 3.5 Sonnet's tool use capabilities Integrates external APIs securely within Snowflake Creates natural conversation flow between user, AI, and data Runs entirely in Snowflake - no additional infrastructure needed

The most interesting aspect is watching Claude determine whether to use the weather tool based on the question context. This implementation just scratches the surface of what's possible with Snowflake Cortex AI and Claude. Consider the possibilities of adding tools for travel recommendations, financial data lookups, or inventory checks. Full quickstart guide: https://quickstarts.snowflake.com/guide/getting-started-with-tool-use-on-cortex-and-anthropic-claude/index.html?index=..%2F..index#0

As we race into an age of everything is fed into GenAI for evaluation or identification how are you preparing your data to make it useful? Does a good output require good input?