Client with security concerns · AI Automation Society

Client with security concerns

I have a client who is looking to have some email triage agents built for a few members of their executive team. This would sort/categorize their incoming emails, route appropriately, and take some actions on specific emails like forwarding internally and/or drafting replies.

I am building the system in n8n and would use a standard LLM (im thinking gpt-4o mini) to handle the "thinking". They were warned by a partner (who admittedly doesnt know much if anything about the automation/AI space) that they shouldn't do it because if they tie their email into anything "open source" they are risking sensitive information leaking.

I've tried alleviating their concerns over a couple phone calls but they dont understand much of what I am discussing. Does anyone have any advice or resources on how to tackle this? How i can show in non technical terms that their information is safe?

Thanks in advance!

5 comments