<aside> ℹ️
For questions, comments, or feedback, please connect with Tony Pelosi on LinkedIn or WhatsApp.
</aside>
This Notion page was used by the judges of the NTUPC x Notion 3.0 Ideathon: Build Better with AI to help evaluate, score, and determine the results.
While this Notion page was helpful for the human judges to use in evaluating and scoring the submissions, it was also used in tandem with Claude Cowork connected to this Notion via the Notion MCP to build an and execute an AI judge. Check out Claude Cowork Chat to see exactly how it created Automated Judging Criteria and Automated Judging Results from the contest details and finalist submissions in Notion.
A new judging criteria was told to us last minute so I did not account for that in the AI judge. Pitch Delivery was added at 10% weight so I just had the AI give everyone at 10 for this criteria. A future enhancement would be real-time (using ambient audio) or near real-time (plugging in the summary from Granola AI) judging of the pitches themselves. Granola AI was also very useful for me to rate each submission by talking out loud as I read each submission and having it summarize my thoughts as you can see in Tony’s Judging Notes.
While individual scores from each judge and total scores have been withheld from public view, the AI judge’s submission has been included in . The AI Judge was surprisingly similar to the human judging, with the same exact #1 and #2, but a different #3.
This has been published as an example of how to leverage Notion MCP with Claude (Chat, Cowork, or Code), how one may reproduce for judging future competitions, and for giving transparency to Ideathon participants on how an AI evaluated their submission as feedback for their next competition.
Student Founder OS by Team Jamuludin
Date our Dates by Team Dates that stick
Context-Aware Learning Workspace by Team Synapse
Unified Capture Engine by Team Decidedly
The Problem Kanban Board by Notion Navigators
Creator Command Center by Deep Roast Oolong Tea