Enhancing and optimizing our HAAS Board Concierge

[TKasperczyk]

Hey team,

First off, I want to extend a heartfelt thanks to @CunninghamTim for their exceptional work on the concierge GPT. However, I've noticed some performance issues linked to its reliance on Bing, as highlighted in the latest video.

I propose a couple of enhancements to address these issues and expand the tool's capabilities:

1. Migrating to GraphQL: This move would offer significant performance improvements and more streamlined interactions at the cost of an additional API authorization step, requiring users to click "Allow" before each request. Despite this, I believe the performance gains justify the shift.

   

2. Enhancing Functionality: By enabling the GPT to delve into commits, pull requests, and navigate the codebase, we could unlock new functionalities. Imagine it helping identify key project areas, creating roadmaps based on repository status, and even collaborating on specific files.
   To showcase these possibilities, I've prepared a PoC available here: https://chat.openai.com/g/g-lIAp9qowx-haas-assistant.
   Check out the capabilities and comparisons with the previous version in this video: https://youtu.be/ThKMfD1x7P4

Additionally, integrating the GPT's definition into our repository could be beneficial. The PR owner would ensure the GPT definition stays updated with new PR submissions from other contributors.

I'm open to feedback on these suggestions. If they don't align with our needs, we can certainly close this discussion.

[daveshap]

Your technological distinctiveness will be added...