Q Insight understands free-form questions — write as if you were asking a fellow analyst. The main rule: clearly state what exactly interests you, and filter the feed before asking.
In short: a good question names the object (brand, product, person, event) and the purpose of the analysis. Bad: "What are these articles about?". Good: "What was written about [person/product/event]?". Refine the answer with follow-up questions in the same chat.
Good and bad questions
| ❌ Too general | ✅ Specific |
|---|---|
| "What are these articles about?" | "What was written about [person/product/event]?" |
| "What's up with the brand?" | "What are the main customer complaints about [brand] on social media?" |
| "Compare everything" | "How are our services discussed compared with [competitor]?" |
Example questions by complexity level
Simple — one aspect
- "What are the main topics in this week's publications?"
- "What do customers criticize most often?"
- "Which positive aspects of the brand are mentioned most often?"
- "Give the main points from mentions where the main topic is [person/event]".
Medium — analysis with context
- "Which sources write about us most actively in a negative way?"
- "How has the tone of publications changed after our latest announcement?"
- "What do users say about the recent changes at [company]?"
- "What do users say about service quality compared with competitors?"
Complex — strategic level
- "Which topics are often mentioned together with our brand — and are any of them risky?"
- "Were there any communication crises involving the company? What were they related to and how did the company respond?"
- "What key narratives are forming around [brand/topic] and who spreads them?"
- "What strategy should we use to improve our reputation on social media?"
How to deepen the analysis through dialogue
Q Insight supports several questions in a row within one chat. Each next question takes the previous conversation into account, so you don't need to repeat the conditions.
Example dialogue:
- "What are the main customer complaints in the mentions?" — Q Insight lists the main groups of complaints.
- "What exactly do they write about delivery delays?" — Q Insight goes into detail on this particular topic, with examples of feedback.
- "How can we fix this, based on the mentions?" — continue the dialogue until you get the depth of analysis you need.
| # | Element | What it is |
|---|---|---|
| 1 | Filters and publication counter | The selection that Q Insight analyzes. |
| 2 | Answer | A structured summary: main points, examples, conclusion. |
| 3 | 👍 / 👎 | Rate the answer as useful or not useful. |
| 4 | Copy | Copy the answer text to the clipboard. |
| 5 | Input field | Write a follow-up question. |
| 6 | Send | Send the question (or press Enter). |
| 7 | Reminder | "ChatGPT can be wrong. Check important information." |
Ready-made templates for typical tasks
| Task | Filters | Question |
|---|---|---|
| Complaint analysis | Brand + "Negative" sentiment | "What are the main customer complaints and which of them occur most often?" |
| Reaction to an event | Brand + period after the event | "How did the media and users react to [event]? What are the main points?" |
| Comparison with a competitor | Competitor keyword | "How are [competitor]'s services discussed? What is praised and what is criticized?" (repeat for your brand) |
| Risk search | Brand + source type | "Are there signs of reputational risks or a crisis in the mentions? What are they related to?" |
| Report preparation | Brand + the period you need | "Give 5 key points from the mentions for this period for a report to management". |
Tips
- Name a specific object: brand, product, person, event.
- Specify the format if it matters: "5 points", "briefly for a manager", "with example quotes".
- One question — one task. Break a complex task into several steps in the dialogue.
- Create a new chat for a new selection or a new topic.
- Check key facts and figures in the feed before using them in reports.
Frequently asked questions
The answer is too general. What should I do?
Narrow down the selection with filters (keyword, sentiment, source type, period) and phrase the question more specifically — with the object name and the purpose of the analysis.
Does Q Insight take previous questions into account?
Yes, within one chat, each next question takes the previous conversation into account.
Where do the percentages and quotes in the answer come from?
Q Insight forms the answer based on the text of the publications in the selection, including the figures and quotes they contain. AI can make mistakes, so check important data in the feed.
If you need more help, write to the support chat or contact your manager.
