Asa Pro (Beta): Recommended Conversation Starters

Start broad: "What are the top 5 Codes in this dataset?"

Then add a filter: "What are the top 5 Codes for <Filter_Name>?"

Then compare: "How do the top Codes differ across <Filter_Name> values?"

Then add time: "How have those Codes shifted over <Time_Period>?"

What are the top 5 Categories (Nets) in the dataset? Include count and percentage.

Show me the top 10 Emotions ranked by frequency.

Which Codes appear most frequently? Provide an executive summary.

What Topics dominate the conversation overall?

What are the top 5 Emotions expressed in each <Filter_Name>?

Compare Topics across different <Filter_Name> values. What stands out?

Which <Filter_Name> has the highest concentration of positive vs. negative emotions?

Break down the most common Nets by <Filter_Name>.

Show me how the top 5 Emotions evolved over <Time_Period>.

Which Codes increased or decreased the most over time?

Compare Topics across <Time_Period> — did priorities shift?

Quantify the difference between the top two Categories.

Provide percentages for each segment.

Show me 3 verbatim examples that illustrate this theme.

Lead your response with an executive summary.

DO start with appropriately-sized datasets – Begin with manageable date ranges and filter combinations. ASA Pro performs best when you're strategic about data scope.

DO ask specific, focused questions – "What are the top 3 themes about customer service in January?" works better than "Tell me everything about customer service."

DO leverage segment comparisons – ASA Pro can compare demographics, time periods, and other filters within your existing analysis. Keep questions focused on a single segment level at a time. Avoid asking about multiple overlapping segments simultaneously.

DO request quantification – Ask for percentages, counts, trends, and rankings. ASA Pro can quantify any aspect of your analysis.

DO ask for verbatim examples – Request specific quotes to illustrate themes. ASA Pro's semantic search will surface the most relevant responses.

DO use natural language – You don't need to memorize exact filter names or theme wording. ASA Pro understands variations and synonyms.

DO build on previous questions – ASA Pro maintains context for up to 10 interactions. Use follow-up questions to drill deeper.

DO verify critical insights – For high-stakes deliverables, cross-check ASA Pro's quantifications against the platform. Trust, but verify.

DON'T use excessively large datasets without testing first – Very wide date ranges + many OE fields + complex filters = higher risk of context window limitations. Start smaller, then expand.

DON'T expect ASA Pro to generate net-new content – ASA Pro analyzes existing Canvs data. It won't create marketing copy or original narratives unconnected to your analysis.

DON'T ask for complex segmentation, not in your dataset – ASA Pro works best with single-level segment analysis. Avoid queries that combine multiple segment layers or ask for segmentation breakdowns on filters that don't exist in your uploaded data.

DON'T ask it to perform platform functions – ASA Pro can't export data, change analysis settings, create new projects, or modify your Canvs configuration.

DON'T assume it remembers unlimited history – Context window is limited to 10 interactions. After that, it may lose earlier conversation threads.

DON'T use vague or overly broad queries – "Tell me about the data" is too open-ended. Be specific about what you want to learn.

DON'T ignore validation warnings – If ASA Pro says it's uncertain or can't find data, there may be a real issue. Don't force it.

DON'T expect it to work outside Canvs data – ASA Pro doesn't have general internet knowledge. It only knows your Canvs analysis.