Available on all CallSine plans and for all targeting types (Auto, Manual, etc).
Overview
The Outreach Agent does not follow a fixed schedule. Once launched, it evaluates performance signals and adjusts how and when it reaches out. These adjustments happen automatically, allowing campaigns to improve over time without manual intervention. The result is a system that becomes more effective the longer it runs.
1. Sequencing: How the Agent Adjusts Cadence
When you create an Agent, you define a baseline cadence on the Strategy page. This sets an initial structure for sending email and LinkedIn outreach.
Once the Agent is live, it begins testing variations of that schedule. These tests allow it to determine which timing patterns result in the highest number of responses.
How it works:
The cadence displayed in the Strategy page is a starting point, not a fixed rule
Timing may shift earlier or later based on results
High-performing timing patterns are used more frequently
Example: If a LinkedIn touch scheduled for Day 3 performs better on Day 4, the Agent gradually shifts more messages to Day 4.
The Agent will not accelerate or batch messages in ways that violate inbox warming or LinkedIn safety limits.
2. Task Prioritization: Deciding What to Send When
Each Agent operates within safe daily send limits. When there are more messages ready than can be safely delivered in a day, the Agent decides what to prioritize based on performance.
Example:
Daily email capacity: 60
Messages scheduled today: 70
Only 60 can be sent
The Agent may focus on net-new messages, follow-ups, or a mix—depending on what is producing the most replies.
If capacity is reached, the remaining messages are not lost. They are held automatically and sent on the next available run.
3. Channel Testing: Finding the Best Order
The cadence shown on the Strategy page might display an initial sequence such as:
Email → LinkedIn → Email → LinkedIn
After launch, the Agent tests different versions to see which perform better for different audiences.
It evaluates:
Whether starting with LinkedIn improves response rates
Whether two LinkedIn touches before email outperform the baseline
Which order is most effective for certain industries or roles
Over time, the Agent shifts prospects toward the most productive order.
4. Content Optimization: Learning From Message Performance
Every message the Agent sends is tracked for effectiveness. Over time, patterns emerge in tone, word choice, call-to-action style, and subject lines.
The Agent monitors:
Which messages receive replies
Which CTAs lead to conversations
Subject line behavior
Differences across industries and job titles
If a particular language style or CTA performs well, the Agent uses it more often for similar prospects.
5. Strategy Matching: Assigning the Right Messaging Framework
When an Agent is created, it generates multiple Content Strategies—different narrative structures for the same persona. These strategies are tested across leads.
As results come in, the Agent:
Identifies which strategies perform best
Matches strategies to similar prospects
Adjusts selections as performance changes
Example: if Strategy A works well for CEOs or Strategy B works well in a particular geography, future outreach is routed accordingly.
You can view Content Strategy performance in the Reporting tab of the Campaign Editor.