Gemini is able to hold a chat thread longer than the other tools I have used. It's also better at pulling relevant context out of other chat threads where it's useful so you don't have to duplicate training the chat with the same pieces of data. And lastly, it is able to modify my content in my voice better than the others, it can be trained to be very much me. Review collected by and hosted on G2.com.

It hallucinates more than I would like. The "Gem" RAGs need a lot more hand holding and QA-ing than other "coded" platforms. It has a tendency to focus on one element of training and not balance priorities, paradox, and unique point of view. e.g. If I tell it to make staffing recommendations on skill and fit first it will make the decision based on cost. It also has acted without permission, and you have to have it "show the work" if you want an accurate data or math answer. Review collected by and hosted on G2.com.