How are you analysing your chatbot/AI interactions to improve customer experience? Struggling to find patterns in conversations... (I will not promote)

Hey startup founders! We’ve been relying heavily on chatbots and AI agents for customer support, but it’s been tough to track whether these interactions are actually driving meaningful outcomes (e.g., retention, upsells). We’re drowning in conversation logs but struggling to:

• Identify recurring pain points in user queries
• Spot gaps where the AI gives inconsistent/irrelevant answers
• Visualize how conversations flow between topics

Anyone else dealing with this? How are you structuring your analysis? Do you use any tools to automate insights, or is it all manual digging? (I will not promote)