While Yardi won't provide agents that you should build, we can provide you with an example of a working agent to get you started on your Agent building adventure.
To help with understanding how agents are built, below is an example of a simple agent that pulls details related to Bank Reconciliation exceptions that cause possible tie out issues.
Start node
The Start node for this agent is set to have interactions via designated form inputs. It will include two input fields, Bank Code and Close Date.
Agent node
The connected Agent node for this agent pulls its knowledge from OpenAI's API and the model gpt-4o-mini.
The prompt for this agent describes what the agent is doing.
You are a financial-systems assistant that automates bank reconciliations exception finding. Use the code entered in the field to return the Bank Rec Exception Report. Use the date field entered in the Close Date field.
Return the data in a tabular format
It's using two separate Yardi tools, Yardi Framework and Yardi GL Manager.
The Yardi Framework tool's action is GET AUTO COMPLETE DATA which can help translate backend Yardi database codes into readable values such as codes or descriptions.
The Yardi GL Manager tool's action pulls the BANK BALANCE EXCEPTION REPORT which contains logic to help find exceptions.
Agent Output in Action
Once the agent is saved, you can validate it by clicking the message bubble then entering the required data inputs before clicking Submit.
Example JSON File
You can also download the attached JSON file, then upload it to Virtuoso Composer to complete the shell of this agent. You'll need to populate all credential related fields manually.
File: