I sometimes hear a misconception that the right language for AI projects is Python, when in reality many AI projects today are built entirely using 3rd party or open source services, like LLMs, and all of the integration happens over standard HTTP API calls. This is something that Ruby is great at, and there any many existing gems that provide the needed functionality.
Operating under this misconception could cause a developer with a Rails app to create a separate service in Python that their Rails app talks to, but that Python app is just making calls to the services that Rails could instead call directly. This level of indirection adds application and hosting cost and complexity.
While there are certainly complex AI tasks that might be better suited for Python, such as creating your own machine learning model, make sure you understand what it is you’re trying to achieve and don’t over-engineer a solution when using a Ruby gem and some API calls will suffice.