What is AI Story Assist 2.0: Asa?
Asa is the AI-powered research sidekick in AI Story Assist 2.0, designed to help you explore datasets in Canvs AI through conversations. Asa enhances your research by providing real-time insights, generating visualizations, comparing data segments, and surfacing verbatim responses—all based on your dataset and its associated filters. Think of Asa as your super-powered research assistant, helping you dive deep into your data and produce actionable takeaways. And, he's fun!
Key Differences Between 1.0 and 2.0
AI Story Assist 1.0
Generates a narrative overview of all responses with quantified key takeaways. Enables users to create narrative summaries for subsets of data, and ask questions about the verbatim responses.
AI Story Assist 2.0
Asa, a conversational AI sidekick that understands your full dataset, including all verbatims and filters, & can build on previous questions to provide deeper insights. It also supports more advanced tasks like visualizations & comparisons across data segments.
Step 1: Enable AI Story Assist 2.0 & Meet Asa
Open any dataset in Canvs AI
Look for the AI Story Assist panel on the right side of the dataset page.
If you currently have 1.0 enabled as your default experience, you will see a banner with a link to "Switch to AI Story Assist 2.0"
If you currently have 2.0 enabled as your default experience, you will see a link to "Start Conversation" to initiate a session with Asa, or click "Load the Previous Session" to continue where you left off.
Note: Users can switch between AI Story Assist 1.0 & 2.0 at any time in the top left corner of the tool, or in your dataset's Viewing Preferences.
AI Story Assist 2.0 Prepares Your Dataset
Canvs will take a moment to load your dataset and Asa will display a welcome message and Conversation Starters when ready.
Asa has access to the verbatim responses, filters, and quantitative data in your dataset from the selected wave(s) and question(s).
Step 2: Ask Asa a Question
Start the conversation! Once your dataset is loaded, you will be greeted with a welcome message and suggested conversation starters. Select one, or go ahead and Ask Asa an initial question about your data!
Recommended topics of conversation:
Link to Ask Asa: Recommended Conversation Starters
Investigate & Summarize:
Ask Asa for a summary of the responses
Ask Asa your research goals
Compare Quant & Qual
Ask Asa about your quantitative data (close-end responses / filters)
Ask Asa about how specific segments responded, and how they compare
Ask Asa to identify drivers of changes in themes over time
Visualize
Ask Asa to create visuals that communicate how different segments responded
Ask Asa to iterate on visuals
Consult
Ask Asa to generate strategic reports or ideation based on responses
Ask Asa to help you identify specific types of example verbatims
Ask Asa what questions you should ask
Ask Asa what visualizations he thinks would be useful for you create
Step 3: Ask Asa a Follow Up (Continue the Conversation)
Continue the conversation. Asa remembers the context of your questions, and what you've already asked or generated within your conversation session.
Recommended follow ups:
Ask for a description or explanation (What does XX mean? Can you explain that theme?)
Iterate on visualizations (Re-order the metrics. Use different colors. Make it easier to read.)
Flag potential misunderstandings or confusion (I'm not sure you got that correct. Can you please try again?)
Dig deeper (Break down responses for specific segments)
Break down complicated questions or tasks into multiple steps (First, identify main themes for each sample. Then, create visualizes for each.)
Step 4: Saving and Exporting Sessions
Asa makes it easy to save and export the insights from your session for future use.
Automatic Session Saving
Asa automatically saves your session tied to the specific dataset until you start a new session or clear it.
Your session is private and can only be accessed by you. Other team members will have their own separate sessions; even if they are having conversations about the same dataset.
Exporting Conversations
You can export your entire session, including insights and visualizations, for further use in reports and presentations.
Copy
Conversation Transcript
Charts and Visualizations
Tables and Data: Export quantitative data as
.csv
or .xlsx files.
Step 5: Troubleshooting and Best Practices
If Asa Doesn’t Recognize a Filter or Segment
Tip: Rephrase your request or check the exact filter name in your dataset. Asa will attempt to match your request to the closest filter but being precise will improve accuracy.
If Asa’s Response Seems Inaccurate
Tip: Point out your concern, and ask for a re-do! Provide more detailed instructions or clarify your question. If Asa gives an uncertain response, you can ask it to flag any assumptions it might have made.
Handling Large Datasets
AI Story Assist 2.0 is currently optimized for datasets <10 MB. If your dataset is too large (based on the size of verbatim response as well as filters), Asa might not be able to load.
Tip: Try selecting fewer waves/questions.
Final Tips for Success with Asa
Be Specific: The clearer your questions, the better Asa’s responses will be. Specify filters, demographic groups, or time periods to get precise insights.
Experiment: Ask follow-up questions or try new prompts to dig deeper into your data. Asa learns from your previous questions in each session.
Trust but Verify: Asa is like a diligent intern—always verify critical insights before using them in your reports.