Hello,
I have built an AI chatbot using n8n with Supabase memory, and I’m facing accuracy issues for scenario and conceptual questions.
I’m looking for a paid 1-hour expert review session to improve retrieval, memory handling, and answer grounding.
Please let me know if this is possible and the cost.
Thanks,
Sathiyakumar
Hey @Sathiyakumar_A Welcome to the n8n community!
You can post paid work on the ‘Jobs’ section, and as for the workflow i am here to assist you with that, as far as you have described your workflow uses a Supabase long term memory and a basic chatbot , what i recommend fixing issues like scenario & conceptual question:
-
Let AI Have ability to use Searches like SerpAPI & HTTP node for more fact based and right answers
-
Try fixing your System Prompt cause most of the times AI does not properly know how to behave to different type of questions asked in your scenarios
-
And re-check memory that does it even fetch the previous data or it is just fetching last few data rows
-
Try Using a better reasoning model like Claude-4.5 for better AI output quality
Let me know if this helps or need any further help with this.