Hello people, I have a POC to testing two real cases with agentforce: as type Service Agent and also, einstein summary tool, to summary messaging sessions. In my test I have changed originals LLM model from prompt builder to another models.
To my service agent, i change LLM model to Claude 3.7 in propmt used at action Answers Questions with Knowledge;
To einstein summary tool, in prompt, I changed to Gemini 2.5 PRO, to generate summary texts about conversation in messaging session.
In yours opinions, which better LLM model to my cases? or which model you are used to Service Agent and Einstein Summary Records.
Thanks!!
Hi Esdras,
Each LLM model has it's own unique way of retrieving data and giving the output.
Claude models work good, GPT latest version as well work well.
You can customize any model using Einstein Studio in Data Cloud and this improves uniqueness of response , reduces repeatable words etc.
For better control its bring your own LLM.
Do let me know in case of further queries.
Thanks