Challenge 3 – Advanced: “The Knowledge Alchemist”

Objective: Build a research assistant that can turn scattered information into useful knowledge.


Requirements:

• Participants must utilize Unified Context Layer only - https://ucl.dev/. See the Unified Context Layer documentation here.
• Participants can use any other platform to build apps, automations, etc but they must call upon UCL
• Connect 3+ data sources (e.g., Google Docs, Jira, Notion, Gmail).
• The AI agent should:
• Search across all sources.
• Summarize or reformat the findings.
• Deliver the output in a single location.

**What does this mean:** You’re building a platform that helps build knowledge. It looks in several places, pulls the right bits together, and returns a tidy, useful result in one spot.
**Typical flow:** 1. User asks a question or gives a topic. 2. Agent queries multiple connected sources, gathers relevant items. 3. Agent clusters and summarizes (e.g., highlights, action items, dates). 4. Agent posts the final brief to one destination (e.g., a Notion page or Slack thread) with links/citations.
**Example ideas:** - A course study guide generator. - A team document finder that gathers info from multiple tools.
**How to embed UCL:** [https://docs.fastn.ai/ucl-unified-context-layer/embedding-ucl-onto-your-ai-agent](https://mlh.link/ghwdata925-fastn-ucl-embed)
**How Unified Context Layer helps:** - Observability and traceability for each query/result to validate your assistant’s output.
**Submission checklist:** - Uses 3+ sources - Produces a concise output (not a raw dump) - Includes simple citations back to source items - optional - Demo shows query, then search, then summary, and the final post - Submit a link to your code repository