AI configuration required. Thinking Partner requires an OpenAI or Gemini API key configured in Settings → AI. See AI Configuration.
What is Thinking Partner?
The Thinking Partner is a persistent AI workspace grounded in your live product data. Unlike a generic chatbot, it has access to everything in your Kansov workspace — every Insight, customer record, Idea, Feature, Goal, and Knowledge Hub entry.
Conversations are organized into named sessions that persist across days and weeks. A session you start on Monday builds on itself through the week — there's no losing context between sessions.
Creating and Managing Sessions
Click New Session and give it a descriptive name (e.g., "Q3 Onboarding Investigation"). Sessions appear in a sidebar list, organized by status. You can search across all sessions by title or content.
What the AI Can Access
The Thinking Partner has access to all data in your workspace: Insights (including customer context and ARR data), Ideas and their scores, Features and their statuses, Goals and OKRs, and Knowledge Hub entries. You can also explicitly link specific items to a session for focused reasoning.
File and Image Attachments
Upload documents, screenshots, or images directly into the chat. The AI will read and reason over the uploaded content alongside your workspace data.
Web Search
When web search is enabled, the AI can search the web to supplement its reasoning with current market data, competitor information, or industry context. Enable in Settings → AI → Web Search.
Creating Items from the Chat
- Click Create Insight on any AI response to turn it into an Insight
- Click Create Idea to convert a suggestion into an Idea
Example Prompts That Work Well
- "What are the most common complaints about the onboarding flow? Summarize the top 3 themes from my Insights."
- "I have 12 Ideas in the roadmap. Given our current Goals, which 3 should we prioritize this quarter and why?"
- "Draft a one-page problem statement for the retention issue, drawing on our Insights and the knowledge hub entry on user lifecycle."
Best practice: Start each session with a clear question. The more specific your opening question, the more useful the AI's reasoning will be.