Skip to main content

Understanding key terms

Understanding key terms (campaign, run, engagement, task, agent)

Written by Nathan Hayman

Quick Answer

The five key concepts in Rivvi are: Campaign (your outreach program), Run (a single execution), Engagement (one patient contact), Task (follow-up action), and Agent (AI that makes the calls). Understanding how these work together is essential to using Rivvi effectively.


Who This Is For

  • New users learning Rivvi terminology

  • Team members who want to understand the system

  • Anyone creating or managing campaigns


What You'll Learn

  • What each key term means

  • How these concepts work together

  • Real-world examples

  • Common terminology questions


The Five Key Concepts

Term

Simple Definition

Example

Campaign

A type of outreach program

"Medication Adherence Reminders"

Run

One execution of a campaign

"January 2025 Diabetes Patients"

Engagement

One patient contact in a run

Call to John Smith on Jan 15

Task

A follow-up action needed

"Call back patient who had questions"

Agent

AI that conducts the calls

"Medication Adherence Agent v2.0"

[INSERT IMAGE: Diagram showing how all 5 concepts relate to each other]


Campaign

What It Is

A Campaign is a reusable template that defines:

  • The purpose of outreach (medication reminders, appointment scheduling, wellness checks)

  • What the AI agent will say and do

  • What automations and follow-ups to create

  • Settings and configuration

Think of it as the "blueprint" for your patient outreach.

Key Characteristics

Reusable: Create once, use many times
Template-based: Start from pre-built templates
Configurable: Customize settings, automations, agents
Persistent: Campaigns exist permanently until deleted

Real-World Examples

Campaign Name

Purpose

Medication Adherence Check-ins

Monthly calls to ensure patients are taking prescriptions

Appointment Reminders

24-hour reminders before scheduled appointments

Annual Wellness Visit Outreach

Invite patients for yearly checkups

Post-Discharge Follow-up

Check on patients 3 days after hospital discharge

Flu Shot Campaign

Promote flu vaccinations during flu season

[INSERT IMAGE: Campaign detail page showing template information]

💡 Tip

Name campaigns by their purpose, not by dates. Good: "Diabetes Medication Adherence". Bad: "January 2025 Campaign".


Run

What It Is

A Run is a single execution of a campaign with:

  • A specific list of patients

  • A specific time period

  • A selected AI agent version

  • A schedule for when calls happen

Think of it as pressing "play" on your campaign for a particular group of patients.

Key Characteristics

One-time execution: Each run happens once
Patient-specific: Contains a defined list of patients
Trackable: Has clear start/end dates and status
Results-oriented: Produces measurable outcomes

Run States

Runs progress through these states:

State

Meaning

What You Can Do

Draft

Being set up, not started

Edit, upload patients, configure

Active

Currently running

Monitor, pause, view progress

Paused

Temporarily stopped

Resume or cancel

Complete

Finished successfully

View results, export data

Failed

Ended due to error

View error logs, retry

Cancelled

Manually stopped

Archive or delete

[INSERT IMAGE: Run state diagram showing progression]

Real-World Examples

Using the "Medication Adherence Check-ins" campaign:

Run Name

Patients

When

Agent

Q1 2025 - Diabetes

500 diabetes patients

Jan 15-31

Med Adherence v2.1

Q2 2025 - Diabetes

485 diabetes patients

Apr 15-30

Med Adherence v2.1

Q1 2025 - Hypertension

300 hypertension patients

Jan 15-31

Med Adherence v2.1

[INSERT IMAGE: Multiple runs listed under one campaign]

📘 Note

You can run the same campaign multiple times with different patients, dates, or settings. Each is a separate "run".


Engagement

What It Is

An Engagement is a single interaction attempt with one patient in a run:

  • One phone call (successful or not)

  • Call duration and outcome

  • Recording and transcript (if call connected)

  • Any data collected during the call

Think of it as "one line in your call log".

Key Characteristics

Patient-specific: One engagement per patient per run
Outcome-tracked: Records what happened
Recorded: Saves call audio and transcript
Data-rich: Captures responses, sentiment, next steps

Engagement Outcomes

Outcome

Meaning

What Happens Next

Success

Call connected, goals achieved

Marked complete

No Answer

Patient didn't pick up

May retry based on settings

Busy

Patient's line was busy

May retry based on settings

Voicemail

Left message

Logged, may create task

Failure

Technical issue or patient hung up

Logged, may create task

Opted Out

Patient requested no more calls

Added to DNC list

[INSERT IMAGE: Engagement detail showing call recording, transcript, and outcome]

Real-World Example

Patient: John Smith

  • Campaign: Medication Adherence Check-ins

  • Run: Q1 2025 - Diabetes

  • Date: January 15, 2025, 2:30 PM

  • Duration: 3 minutes 42 seconds

  • Outcome: Success

  • Data Collected: Confirmed taking medication, no side effects, refill scheduled

  • Tasks Created: None (patient compliant)


Task

What It Is

A Task is a follow-up action that requires human attention:

  • Call back a patient who had questions

  • Reschedule an appointment

  • Verify medication availability

  • Address a concern raised during the call

Tasks are automatically created based on campaign automations or call outcomes.

Key Characteristics

Action-oriented: Requires someone to do something
Assigned: Can be assigned to specific team members
Trackable: Has status (Open, Resolved)
Linked: Connected to specific patient and engagement

Task Types

Type

Created When

Example

Callback Required

Patient couldn't talk, requested callback

"Call Mary Johnson back at 3pm"

Question/Concern

Patient had a question agent couldn't answer

"Patient asked about medication dosage"

Appointment Needed

Patient needs to schedule something

"Schedule colonoscopy for James Lee"

Verification Needed

Information needs human verification

"Verify pharmacy has prescription"

Custom

Based on your campaign automations

Anything you configure

[INSERT IMAGE: Task list showing various task types and statuses]

Task Statuses

Status

Meaning

Open

Not yet addressed

In Progress

Someone is working on it

Resolved

Completed successfully

Failed

Couldn't be completed

Cancelled

No longer needed

Real-World Example

Task: Callback Required

  • Patient: Sarah Williams

  • Campaign: Medication Adherence Check-ins

  • Run: Q1 2025 - Diabetes

  • Created: January 15, 2025, 2:45 PM

  • Reason: Patient was at work, requested callback after 5pm

  • Assigned to: Maria (Care Coordinator)

  • Status: Resolved

  • Resolution: Called back at 5:30pm, patient confirmed medication adherence

[INSERT IMAGE: Task detail view showing all information]


Agent

What It Is

An Agent is the AI voice that:

  • Calls patients

  • Holds conversations

  • Follows scripts and logic

  • Collects information

  • Creates tasks when needed

Think of it as a very smart, never-tired team member who makes calls.

Key Characteristics

Voice-based: Sounds natural and human-like
Conversational: Can handle questions and variation
Purpose-built: Each agent designed for specific use cases
Versioned: Agents improve over time with new versions
HIPAA-compliant: Secure and privacy-focused

Agent Versions

Agents are continuously improved. You can select which version to use:

Version

What Changed

When to Use

v1.0

Original release

Testing only

v1.5

Improved question handling

Stable, proven

v2.0

Natural language updates

Recommended

v2.1

Spanish support added

Latest features

💡 Tip

Always use the latest recommended version unless you have a specific reason not to. New versions are better at handling complex conversations.

[INSERT IMAGE: Agent selection dropdown showing versions]

What Agents Can Do

✅ Introduce themselves and explain purpose
✅ Ask questions and listen to responses
✅ Confirm appointments or medication adherence
✅ Schedule follow-ups
✅ Transfer to human if needed
✅ Understand and respond to common questions
✅ Detect sentiment (frustrated, confused, etc.)
✅ Create tasks for human follow-up

What Agents Cannot Do

❌ Diagnose medical conditions
❌ Provide medical advice
❌ Prescribe medications
❌ Handle emergency situations (will transfer)
❌ Make complex medical decisions


How They All Work Together

Let's follow a complete example:

[INSERT IMAGE: Flow diagram showing complete process]

Example: Medication Adherence Program

Step 1: Create Campaign

  • Name: "Monthly Diabetes Medication Check-ins"

  • Purpose: Ensure patients take insulin regularly

  • Settings: Create task if patient reports issues

Step 2: Create Run

  • Name: "January 2025 - East Clinic Patients"

  • Patients: Upload 200 diabetes patients from East Clinic

  • Agent: Select "Medication Adherence v2.1"

  • Schedule: Calls between Jan 15-20, 9am-5pm

Step 3: Launch Run

  • Run starts automatically on Jan 15 at 9am

  • Agents begin calling patients

Step 4: Engagements Happen

  • Engagement 1: John Smith

    • Outcome: Success

    • Data: Taking medication correctly

    • Tasks: None created

  • Engagement 2: Mary Johnson

    • Outcome: Success

    • Data: Ran out of medication

    • Tasks: "Contact pharmacy to verify refill"

Step 5: Tasks Created

  • Task for Mary Johnson assigned to care coordinator

  • Care coordinator calls pharmacy

  • Task resolved

Step 6: Run Completes

  • All 200 patients contacted

  • Results:

    • 150 successful, medication adherence confirmed

    • 30 no answers (will retry)

    • 15 tasks created for follow-up

    • 5 requested callback


Common Questions

Q: Can I have multiple runs of the same campaign running at once?

A: Yes! You can run the same campaign multiple times simultaneously with different patients.

Q: What's the difference between a campaign and a run?

A: A campaign is the template/blueprint. A run is executing that template with actual patients. Think of it like a recipe (campaign) vs actually cooking the meal (run).

Q: How many engagements are in a run?

A: One engagement per patient. If you upload 500 patients, you'll have 500 engagements.

Q: Do tasks expire?

A: No, but you should resolve them promptly. Use filters to find old open tasks.

Q: Can I change agents mid-run?

A: No. Once a run starts, you cannot change the agent. Create a new run if you need a different agent.

Q: What happens if a patient is in multiple runs?

A: They'll receive calls from each run independently. Make sure you're not over-calling patients by checking their engagement history.

Q: Can one task belong to multiple patients?

A: No. Each task is patient-specific and links to one engagement.


Visual Reference

The Hierarchy

Organization └── Campaign     └── Run         └── Engagement (per patient)             └── Task (if needed)

[INSERT IMAGE: Hierarchy diagram showing relationships]

The Flow

1. Create Campaign (template) 
2. Create Run (select patients, agent, schedule)
3. Launch Run
4. Engagements happen (calls made)
5. Tasks created (when follow-up needed)
6. Tasks resolved (team addresses them)
7. Run completes
8. Review results

[INSERT IMAGE: Flow diagram with icons]


Next Steps

Now that you understand these terms:

Did this answer your question?