Below is the task log from the launch of my prototype- a Stakeholder Relations Management agent. Finally launched it yesterday.
| Breyden | Breyden’s AI agent |
|---|---|
| Assigns the SRM agent to the contact by tagging the client. | Retrieves contact/situation specific info from the database and initializes a ‘starter state’ |
| Calls the contact | |
| Explains the reason for her call | |
| Requests preferences/availability from the contact | |
| Compares availability | |
| Finds a slot that works for rescheduling | |
| Fields any questions from the contact | |
| Wraps up the call | |
| Books the appointment in Breyden's calendar | |
| Emails Breyden requesting confirmation of his availability for the reschedule slot and waits to receive | |
| Confirms the appointment by accepting in email or calendar | Once confirmed internally, texts the contact asking them to confirm |
| Creates a Google meet event | |
| Sends the link to Breyden’s email | |
| Updates the calendar event with the link | |
| Emails the event and link to the contact | |
| Saves the call as an audio file | |
| Adds a link to the audio file and a full transcript of the call to the Notes section of the contact's profile in the database | |
| Initiates a new ‘flow state’. Further updates the database with any other important info about the contact | |
| Texts/Emails the contact the day before the appointment, an hour before the appointment, and while Breyden is “going online now" in Google Meet at start time | |
| Reminds Breyden of the appointment as well | |
| Updates stakeholder status based on meeting results/attendance | |
| What a freeloader… | Chooses next path (nurture, learning campaign, move to not interested, etc.) |
Despite some user experience hiccups I'm ironing out, The SRM agent completed all tasks accurately.