I run a coaching SaaS on HighLevel where each student has their own sub-account. Inside their sub-account, they create structured “digital assets” — their USP, avatar, brand promise, offers, etc. These assets are unique per student and must stay isolated within their sub-account.

I want the ability to build custom agents using Agent Studio inside my own agency sub-account (e.g., a test/ops sub-account) and then make those agents available to my students through Ask AI across all their sub-accounts.

The workflow I’m trying to enable:

Setup

I create an agent (e.g., a “Blog Creator Agent”) in my main agency sub-account.

I expose this agent inside Ask AI so students in other sub-accounts can use it.

Each student’s sub-account contains their own digital assets and contextual data.

Runtime

When a student in Subaccount B asks Ask AI:

“Create a blog post for my offer,”

the system should:

Route the query to my agent stored in Subaccount A.

Allow the agent to securely access the student's data in Subaccount B (USP, avatar, offer, brand voice, etc.).

Generate a personalized output using that sub-account's data.

Return the result to Ask AI inside the student’s sub-account — without exposing my agent’s internal logic or prompts.

In short:

Agent logic lives in the agency's sub-account (A).

Agents should be able to securely read student data from student sub-accounts (B).

Students only see the final result, not the agent’s backend logic.

This would enable agencies like mine to create reusable, proprietary AI agents that provide personalized outputs for each student—powered by the student’s own data but centrally maintained by the agency."

An extra feature to this would be that we can add custom fields to the brand voice "form". That way we can "inject" our methodology framework calling on those fields contents through the agents.