Gathering Feedback

AI assistant should proactively surface relevant marketplace apps, when native features don't solve user's problem

Description: The monday.com AI assistant currently has two gaps when handling feature requests that could be solved by marketplace apps:

Gap 1: Marketplace apps are not proactively suggested

When a user asks the AI how to accomplish something that isn't natively supported, the assistant responds with manual workarounds only — without mentioning that marketplace apps exist that solve this problem directly. The user must then follow up with an explicit second question like "are there any apps f

or this?" to get app recommendations at all.

This creates unnecessary friction and means many users never discover relevant apps — simply because they didn't know to ask a second question.

Requested behavior: When the AI identifies that a user's goal isn't fully achievable natively, it should proactively add something like: "There are also marketplace apps that solve this directly — would you like me to suggest some?"


Gap 2: App recommendation ranking doesn't reflect marketplace quality signals

When the AI does recommend apps, the ranking appears to be based on community forum mentions and indexed web content — not actual marketplace data. This means well-established apps with strong install bases and high ratings can be overlooked in favor of apps that happen to be more discussed online.

Requested behavior: Incorporate structured marketplace signals into app scoring, including:

  • Install count

  • Average review score

  • Number of reviews (to weight confidence)

  • Recency / last updated


Why this matters:

Users trust the AI to guide them to the best solution. Right now it's failing twice — first by not mentioning apps exist, then by not ranking them well. Fixing both would meaningfully improve user outcomes and create a fairer, merit-based environment for marketplace developers.