What is the challenge? |
I want to schedule an AI agent to run autonomously, based on a defined frequency. Alternatively I may want to prompt an agent to run using an automation. The challenge is that changes should be "co-authored" by the AI agent and an Aha! user. The AI Assistant should not be allowed to make changes (add/update/delete records) without an Aha! user reviewing them first. |
What is the impact? |
Users must initiate AI agent actions. There is not a way to get certain agents to run automatically. |
Describe your idea |
Allow users to prompt agents in two new ways:
Prompt the user when there are AI actions that need to be reviewed/approved. Options:
|
I'm completely behind this approach and suggestion. Great minds think alike. Having a way for the agent to execute certain tasks on a schedule will allow users to take care of SO MANY actions/automations currently unavailable in Aha. I would not be surprised if a significant number of Future Consideration requests on this portal can just be handled through this approach. Here are a few quick examples that come to mind:
1) Release Phases and their To-do items (created through a release template) are not available through Automations, no problem. Write a prompt to review a specific report which will provide a set of IDs for it to update (e.g. setting assigned users based on Workspace names, custom fields etc.)
2) Manual process to link up all Features under Master Feature to higher levels in object structure, no problem. Schedule an AI script to periodically look up whether things are linked up correctly and execute.
3) No way to set additional required fields post object creation on later states, no problem. Schedule an AI agent job to run through a report identified ID list and do it.
4) We're dealing with 20 different workspace integration configuration and want to confirm that all the fields are set up according to a specific template, but we can't use the Aha integration template because a few of the 20,30,40 fields have to be different. Let scheduled AI assistant handle it.
5) Governance of field population would be a breeze to set up.
PS: Approval on certain executions is an excellent idea.
I think the sky is the limit on this one especially due to the fact of how Aha SW is structured.