Create and Monitor Your Agentic Campaigns
Learning Objectives
After completing this unit, you’ll be able to:
- Explain the steps to create an agentic campaign.
- Describe how the agent defines an audience and generates a campaign plan.
- Identify the key sections of the Optimization Insights tab and how to interpret campaign performance data.
- Summarize how the agent rebalances channel allocation to optimize toward a goal.
Launch Isabelle’s Campaign
With the Marketing Goals Agent configured, Isabelle Givens is ready to put it to work. Northern Trail Outfitters is launching the Urban Explorer Plan 2026—a subscription renewal campaign targeting lapsed and at-risk subscribers across North America. Isabelle’s goal is a 25% renewal lift over the course of the campaign, with a $500,000 message budget, split across email and SMS. She doesn’t want to spend days configuring campaign logic. She wants to describe what she’s trying to accomplish and let the agent figure out the rest.
Let’s help her get there.
Create an Agentic Campaign
From the Marketing app, click the Campaigns tab and click Build Agentic Campaign. Enter a campaign name and select the Business Unit where you want to run the campaign—all downstream content, segments, and flows are scoped to the business unit you choose here. In the natural language prompt, describe the campaign objective in plain terms: what you’re trying to achieve, your message budget, preferred channels, and any program context. In Isabelle’s case, it would be:
“Grow NTO subscription renewals by 25% by targeting lapsed and at-risk subscribers across North America. Total message budget: $500,000. Use email and SMS. Parent program: NTO Customer Retention.”
Once you’ve described your objective, select the channels you want to use and click Generate Plan. From there, set your campaign duration, then choose your goal.
Define Audience, Content, and Send for Approval
Add Content Variants and Define Segments
Content variants give the agent something to test against—options it can choose from and rebalance across as real-time performance data comes in.
To set this up, select a workspace–either CMS or the Agentforce Content Agent workspace, browse through your published variants, and add whichever ones you want the agent to test. Keep in mind content assets need to be in Published state to show up here, so make sure you’ve published everything before kicking off the campaign.
At NTO, Isabelle has already worked with her content team to prepare several email and SMS variants for the renewal campaign—different subject lines, offers, and CTAs aimed at different subscriber motivations. She adds all of them here so the agent has a range of options to test against.
From there, you just need to define your audience. You can either let the agent build your target audience automatically using Data 360 segments, or take manual control yourself.
Option | Outcome |
|---|---|
Generate Segment (default) | The agent uses Data 360 to identify the optimal audience, and shows you which populations are included and excluded. |
Existing Segment | You select your own published segment—use this when you have specific targeting criteria the agent shouldn’t override. |
Isabelle decides to let the agent identify her optimal audience.
Review and Send for Approval
Review your campaign summary. When you’re satisfied, click Send for Approval. This kicks off the approval flow you configured in Unit 2, which routes the execution plan to the right approver.
Once it’s approved, the agent takes over, creating and publishing the campaign on its own. There’s nothing more you need to do to launch.
Monitor Performance in Optimization Insights
After your campaign is live, head to the Optimization Insights tab to track how it’s performing. Open the campaign from the Agentic Campaigns tab and click Monitor Campaign.
Section | What It Shows |
|---|---|
Agentic Summary | AI-generated plain-language health check of your campaign, including the MGA Success Forecast Score (predicted likelihood of success) |
Key Metrics | Four KPI tiles: Goal Achievement, Message Budget Spent, Cost Per Goal Unit, and Audience Reached—each with an On Track or Off Track status |
Goal Achievement versus Forecast | Actual progress against projected performance, with On Track / At Risk / Exceeding status |
Message Budget Spend Trend | Cumulative actual spend vs. forecast projection, with your total message budget limit shown |
Channel Efficiency | Email CTR (click-through rate) and SMS CTR vs. baseline—performance above baseline indicates strong engagement |
Rebalance | Allocation weight over time by channel and content variant, plus a log of rebalance events with dates and decisions |
A few days into the campaign, Isabelle checks in on the Agentic Campaign Insights. The agent has already made two rebalancing decisions—shifting more message budget toward email after SMS click rates came in below baseline. She doesn’t need to intervene; she just needs to know the campaign is on track.
Read the Agentic Campaign Insights Like a Pro
The Agentic Campaign Insights gives you a lot of data points. Here’s how to make sense of them without getting lost.
- Focus on trends, not single snapshots. Individual data points fluctuate. Look for sustained patterns over multiple days before drawing conclusions.
- Cross-reference your metrics. No single number tells the full story:
- High spend + low goal achievement = poor efficiency
- Low audience reach + high spend = delivery or targeting issue
- High CTR + low conversions = engagement without conversion
- High spend + low goal achievement = poor efficiency
- Check Rebalance Events when performance shifts. When you see an unexpected change in spend, channel mix, or content performance, open the Rebalance section and select the date in question. The agent logs every rebalancing decision it made and why.
Wrap Up
You’ve come full circle—from understanding the problem agentic campaigns solve, to setting it up as an admin, to walking through a full campaign creation and monitoring workflow as a marketer.
At its core, the agent changes how your team works. Instead of manually managing campaign logic, you’re defining strategic intent and letting AI handle execution. Human oversight is built in at every step through approvals and the Agentic Campaign Insights. It’s a fundamentally different way to run campaigns—and now you know how to make it work.
