Running Campaigns

Execute and monitor your marketing campaigns with AI orchestration

Running Campaigns

Once you've created a campaign, running it launches the autonomous AI Campaign Orchestrator—a sophisticated agent that executes all your campaign deliverables in an intelligent order without requiring manual intervention. The orchestrator understands dependencies between deliverables (research informs content, content guides creative assets), monitors quality at each step, and adapts its approach based on what's been created so far. This means you can set a campaign running and check back later to find all deliverables completed, or monitor progress in real-time if you prefer.

Starting Campaign Execution

To run a campaign, you have two options: use the Voice AI Agent or the dashboard interface. Via voice, simply say "Run the [campaign name] campaign" and the agent will find the matching campaign and start execution. If you have multiple campaigns with similar names, the agent will ask for clarification or default to the most recent one. Via the dashboard, navigate to the Campaigns page, find your campaign card, and click the "Run" button. The campaign status will change from DRAFT or READY to RUNNING, and you'll see progress updates as deliverables complete.

When a campaign starts running, TractionDesk creates a dedicated AI agent session with full context about your campaign. This session includes your campaign's name, topic, objective, selected deliverables, and—crucially—your organization's brand context from the RAG knowledge base. The orchestrator agent uses a sophisticated instruction set that guides it through deliverable execution, quality checking, and decision-making. For example, if your campaign includes both images and copy, the orchestrator will generate the copy first, then create images that visually complement that copy.

The Campaign Orchestrator follows a standard execution order to maximize coherence across deliverables. First, it creates and saves a campaign plan that outlines key messages, specific actions, and a suggested timeline. If research is included, that runs next since findings inform all subsequent content. Then written content (blog posts, social posts, email copy) is generated, followed by visual assets (images and videos) that reference the written content for consistency. This intelligent ordering ensures that later deliverables can build upon and enhance earlier ones rather than being created in isolation.

Monitoring Progress

While a campaign runs, you can monitor its progress in real-time on the Campaigns page. Each campaign card displays the current status and shows which deliverables have completed versus which are still pending. Click on a campaign card to expand the detailed view, where you'll see individual campaign steps with timestamps for when each started and completed. This granular visibility helps you understand exactly what the AI is doing at any moment.

The detailed campaign view also provides links to generated assets as they're created. When an image deliverable completes, you'll see a thumbnail and can click through to view the full image. Copy deliverables show text previews. Video deliverables display the video player once generation finishes. This immediate access means you can review content while the campaign is still running and decide whether to use it or regenerate if you're not satisfied.

If a campaign fails mid-execution, the status changes to FAILED and you'll see an error message explaining what went wrong. Common failure reasons include insufficient credits (the campaign tried to generate a video but you didn't have the required 20 credits) or temporary API issues with content generation services. When a failure occurs, the campaign preserves all completed deliverables—you won't lose work that was already done. Simply address the issue (add credits, wait for API recovery) and re-run the campaign. The orchestrator is smart enough to skip deliverables that already completed and resume from the failure point.

Campaign Execution Time

Campaign execution time varies based on the number and type of deliverables. Simple campaigns with just copy generation might complete in 30-60 seconds. Medium complexity campaigns with research, copy, and a few images typically take 3-5 minutes. Large campaigns with multiple videos, extensive research, and numerous creative assets can take 10-20 minutes. Video generation is the slowest deliverable (60-120 seconds per video), while copy and images are faster (10-30 seconds each). Research duration depends on query complexity and number of sources scraped.

The Campaign Orchestrator runs deliverables in parallel when possible to minimize total execution time. For example, if your campaign includes 5 images, the orchestrator may request all 5 simultaneously rather than sequentially, cutting the total image generation time from 5 minutes to 1 minute. However, deliverables with dependencies (like creating social posts that reference a blog post) must run sequentially to ensure coherence.

Campaign Results and Assets

When a campaign completes, all generated assets are saved to your account and organized by campaign. You can access them through the campaign detail view or through their respective sections (Images, Copy, Videos, Research). Each asset includes metadata about which campaign created it, making it easy to track ROI and understand what content came from which marketing initiative.

Generated content is ready to use immediately—you can download images, copy text to your clipboard, share video links, or export research reports. Many users create campaigns specifically to build a content library, running a campaign that generates 20-30 pieces of content at once, then drawing from that library over weeks or months for social media posts, blog articles, or ad creative. This batch approach is more efficient than generating content piece by piece and ensures thematic consistency across your marketing efforts.