Campaign Optimizer Agent – FAQs

Variants are typically ready within 20–30 minutes of enabling the Campaign Optimizer Agent on a trigger.

You'll receive an email notification once the variants are generated and ready for review.

Yes. In addition to the suggested variants (A–C), you’ll see a Variant D (Custom) option. With Variant D, you can:

• Copy Variant A/B/C, edit it, and save
• Pick from the Content Library — a collection of all variants generated by the tournament
• Start fresh and write your own subject/preheader

Currently, optimization is limited to the subject line and preheader only.

If the test underperforms significantly, it's archived, and the control variant continues. The Campaign Optimizer Agent may then generate a new variant to test.

Yes. You can stop the test at any time. This will route all traffic to the control variant and restart the content generation process.

Each trigger is treated independently. The Campaign Optimizer Agent must be enabled per trigger, and variants are created separately for each trigger.

• Generating content: The Campaign Optimizer Agent is creating and ranking test variants for approval.
• Approval pending: Variants are awaiting user approval before testing can start.
• Approval denied: Variants were rejected during approval; this set will not proceed to testing.
• Approval expired: If the test was in Generating content or Approval pending status and you switched the mode from Optimizer Agent to Manual, the status changes to Approval expired. This ends the approval process.
• Test running: An approved test is in progress (optimization or observation).
• Success: The variant achieved significantly positive lift and is promoted over the control.
• Failure: The variant did not meet success criteria (negative lift or no significance within the limit).

• Campaign Optimizer Agent: AI engine for discovery, content generation, and testing.
• Opportunity discovery pipeline: List of surfaced optimization suggestions.
• Opportunity score: Combined metric measuring potential for improvement.
• Content score (AIDA): Quality score based on Attention, Interest, Desire, Action.
• Priority score (RICE): Prioritization metric based on Reach, Impact, Confidence, and Effort.
• Tournament history: Record of content generation tournaments. Winning variants from these tournaments are proposed by the Campaign Optimizer Agent for A/B testing (not a history of A/B tests themselves).
• Lift: Observed improvement in performance between control and test (projections are not shown).
• Statistical significance (sig): Confidence level that results are not due to chance.
• Traffic allocation: Percentage split of user traffic between control and variants.
• Conversion rate: Percentage of users completing the desired action.

The Pending Actions tab is the main entry point for system-suggested tests. It lists high-potential triggers ranked by score, with rows showing one of two statuses:

• Approval pending: Variants already generated and waiting for your review.
• Opportunity: A system-surfaced suggestion where variants have not yet been generated.

Clicking Discover Opportunities refreshes the list by scanning campaigns and triggers in real-time to identify new optimization suggestions.

No. Clicking Content Generation only creates variants you can review and set up for testing. The test itself does not begin until a variant is approved and launched.

The end date depends on the send volume of the trigger in the last 7 days. The Campaign Optimizer Agent estimates how long it will take to reach the minimum sample size needed for a stable, statistically significant result. For low-volume triggers, this duration can extend up to 90 days.

A test may show significantly positive lift early, but the system continues until the optimization phase ends. This helps rule out initial bias or variance. The optimization phase length is based on the days required to reach the minimum sample size from the trigger’s send volume. If needed, you can manually copy the test variant to the control and end the test once you’re confident in the results.

Links in AI-optimized triggers from the Campaign Optimizer Agent include two tracking parameters:

• bsft_variant_src — Indicates whether the variant was AI-generated (ai) or human-authored (non-ai).
• bsft_variant_type — Indicates whether the variant was the control (control) or the test (test) at the time of sending.

These parameters appear only in the message links. They do not appear in the Campaign Optimizer Agent interface or reports.