How to do Voice of Customer in the Age of AI: A Case Study From My Own Build

Source: Nate’s Newsletter Date: 2025-05-16 URL: https://natesnewsletter.substack.com/p/how-to-do-voice-of-customer-in-the

Summary

Nate documents how he used o3 to analyze anonymized customer feedback data for his PromptKit product launch, collapsing what would traditionally be hours or days of VoC research into minutes. The case study shows AI being applied not just to code generation (60% of PromptKit’s codebase) but to the research and prioritization layer — using LLM analysis to narrow MVP scope based on real signal. The method: collect, anonymize, and batch-analyze customer data through a reasoning model, then let the output drive product decisions.

Implications

• VoC research is becoming a commodity operation for AI-fluent builders; teams that still treat it as a manual, time-intensive practice are operating at a structural speed disadvantage.
• The pattern of using o3 (or equivalent) for synthesis over raw customer data is likely to generalize broadly — any qualitative corpus (support tickets, interviews, reviews) is now tractable at speed.
• Feeds thread: AI-accelerated product discovery / solo and small-team building patterns.