Overview
Spark Thinking Orchestration gives you visibility into how Spark builds an answer.
When you ask a question, Spark may use multiple specialized AI assistants and data tools to gather information, analyze results, and generate insights. Spark Thinking shows this process in real time so you can better understand how your answer was created.
This added transparency helps you feel more confident in the insights Spark provides—especially for more complex analysis.
Why This Feature Exists
AI-powered answers can sometimes feel like a “black box.” You see the result, but not how it was created.
Spark Thinking changes that by showing:
- Which assistants were involved
- When data or tools were used
- The key steps taken to build the answer
This makes it easier to trust the results and understand how Spark is working on your behalf.
How It Works
When you submit a question, Spark may break it into smaller tasks and assign them to different specialized assistants.
These assistants may focus on areas like:
- Market intelligence
- Competitive analysis
- Messaging analysis
- Data retrieval
- Insight synthesis
Spark brings all of this together into a single, unified response.
As this happens, the Spark Thinking panel updates in real time to show the progress.
The Spark Thinking Panel
The Spark Thinking panel appears alongside your answer and updates as Spark works.
Here’s what you’ll see:
Assistants Used
Each assistant represents a specific capability Spark used to complete your request.
Examples include:
- Competitive analysis
- Messaging analysis
- Trend analysis
Status Updates
Assistants will move through simple states such as:
- Working
- Complete
This helps you understand when each step has finished contributing to your answer.
Tool Usage
Some assistants may use internal tools or data sources.
When this happens, Spark Thinking will indicate that tools were used as part of the analysis.
Example
For a question like:
“Compare my competitors on messaging themes this quarter.”
Spark may:
- Gather relevant content and coverage
- Analyze messaging themes across brands
- Compare positioning between competitors
- Combine findings into a summary
You’ll see these steps reflected in the Spark Thinking panel as the answer is being generated.
What You’ll See (and What You Won’t)
Spark Thinking is designed to provide helpful transparency while protecting system integrity.
You Will See
- Which assistants were used
- A high-level view of the steps taken
- When each step completes
You Won’t See
- Internal system prompts
- Raw model outputs
- Proprietary algorithms
- Sensitive internal data queries
This ensures you get meaningful visibility without exposing technical details.
Why It Matters
Spark Thinking helps you:
Build Trust
See how your answer was created and feel more confident in the results.
Understand the Process
Gain insight into how Spark analyzes and combines information.
Ask Better Questions
Use what you see to refine your queries and explore deeper insights.
Best Practices
To get the most out of Spark Thinking:
Ask Analytical Questions
This feature is most helpful when Spark is doing multi-step analysis.
Examples:
- “Compare our brand perception against competitors this quarter.”
- “Summarize messaging trends in recent coverage.”
Use Follow-Up Questions
If you see a step you want to explore further, ask a follow-up.
Example:
“Can you go deeper into the messaging themes you identified?”
FAQs
Will every question use multiple assistants?
Not always. Simpler questions may only require one step.
More complex questions will typically involve multiple assistants.
Why do steps sometimes appear in a different order?
Spark works dynamically, and some steps may complete faster than others depending on the task.
Can I choose which assistants Spark uses?
No. Spark automatically determines the best way to answer your question.
Summary
Spark Thinking gives you a clearer view into how your answers are created.
By showing the assistants and steps involved, it helps you:
- Understand how insights are generated
- Build confidence in AI-driven analysis
- Work more effectively with Spark