Sendō Workflows: Video Creative Optimization

Overview

Sendō is an agentic guide integrated directly into the junbi platform. Powered by a multi-agent system, Sendō combines neuroscientific research, visual asset recognition, and video performance data to help you analyze specific creative assets before launch.

Strategic value of prompt-driven analysis

Ad-level summary scores show overall performance, but prompt-driven analysis explains why those scores occur. By asking Sendō targeted questions about your Metrics over time, you transform diagnostic data into precise visual evaluations that validate your creative hypotheses.

Step-by-step guide

Step 1: Open Sendō on the Prediction Page

1. Log in to junbi and select your project.
2. Open the Prediction Page for the video asset you want to evaluate.
3. Click the blue Analyze with Sendō button in the bottom right corner of the screen.
4. The Sendō sidebar panel will open on the right side of the screen.   

Step 2: Upgrade from base prompts to Masterclass prompts

When you open the panel, Sendō displays four quick-start base prompts:

• What is my brand attention score?
• How can I improve my cognitive ease in my video?
• What is cognitive ease?
• Analyze my video performance

Base prompts provide a helpful summary. However, Masterclass prompts unlock second-by-second diagnostic insights.

Step 3: Apply the Context + Timeframe + Action formula

To receive exact frame analysis, structure your custom prompt using three core components:

1. Context: Specify the diagnostic score you want to evaluate (Brand Attention, Ad Breakthrough, or Cognitive Ease).
2. Timeframe: Define an exact window using Metrics over time data.
3. Action: State what you want Sendō to diagnose or evaluate.

Comparing a base prompt to a Masterclass prompt

• Base Prompt: How can I improve my cognitive ease in my video?
• Masterclass Prompt: "Analyze the metrics over time for Cognitive Ease between second 8 and second 12. What specific visual elements cause visual clutter in these frames?"

When you provide this level of detail, Sendō evaluates frame transitions, tracks primary focal points, and analyzes background complexity to explain score fluctuations.

Step 4: Use core prompting frameworks

Framework 1: Diagnostic prompts (Finding the root cause)

Use diagnostic prompts to pinpoint exact visual friction points in a scene.

• Masterclass Prompt: "My Cognitive Ease score dips sharply around second 09. What visual elements are competing for attention in this scene, and how does the composition cause visual clutter?"

Framework 2: Strategic prompts (Understanding commercial impact)

Use strategic prompts to connect performance curves with visual retention principles.

• Masterclass Prompt: "How does the balance between early Ad Breakthrough and Cognitive Ease in the first 3 seconds impact viewer attention based on alpha.one research?"

Framework 3: Hypothesis validation prompts (Testing planned edits)

Use hypothesis validation to test intended video edits against performance curves before spending budget on re-editing.

• Masterclass Prompt: "We plan to cut the party scene between second 08 and second 11 to shorten the video. Based on the Cognitive Ease and Brand Attention curves, how will removing this segment impact the viewer flow?"

💡 Pro Tip:

Focus on asking "How" or "Why" rather than simple "What" questions. Instead of asking Sendō "Is my score good?", ask "My Cognitive Ease drops between second 08 and second 12. How does the visual layout in those frames affect cognitive load?"