In a recent LinkedIn post, Teresa Torres shares evolving perspectives on the use of Artificial Intelligence (AI) in analyzing customer interviews. Initially skeptical, Torres, in conversation with Petra Wille, revisits the topic after conducting several experiments, offering a nuanced view on AI’s capabilities and limitations in continuous discovery processes.
Shifting Stance on AI-Powered Synthesis
Torres highlights a significant shift in her thinking regarding AI’s role in processing customer interview data. While acknowledging initial resistance, her recent experiences have led to a more balanced understanding. She notes that the effectiveness of AI tools is heavily dependent on the quality of the input and the user’s expertise.
“AI isn’t magic. It can help, but only if your interviews are strong and you provide the right context.”
This statement underscores a core theme: AI is an assistive tool, not a replacement for fundamental skills. According to Torres, the quality of customer understanding, which she terms a “competitive moat,” can be significantly eroded if teams rely too heavily on AI without possessing solid interviewing and analytical foundations.
AI: Accelerating Experts, Lifting Beginners
The discussion delves into how AI impacts different levels of user expertise. Torres points out that while AI can “raise the floor for beginners,” its most significant benefits are realized by experts. This accelerated performance for experienced professionals is a key finding from their experiments.
Torres explains that the real performance gains come from the synergy between an expert user and AI. However, she cautions against viewing AI as a shortcut for the complex process of synthesis.
The Indispensable Human Element
A crucial aspect of Torres’s analysis revolves around the distinction between analysis and synthesis, and the irreplaceable nature of human empathy in customer understanding.
“You still need to synthesize every interview individually. Dumping transcripts into an LLM isn’t a shortcut.”
As Torres argues, the act of synthesizing interview data requires a deep, human-driven understanding that AI, in its current form, cannot replicate. Empathy, she stresses, is a product of direct human interaction, a crucial component that outsourcing to AI would jeopardize.
Navigating the Trade-offs
The conversation, as detailed in the post, explores the practical trade-offs teams face when integrating AI into their discovery workflows. Torres shares personal anecdotes, including a detour into PC building, to illustrate how even complex tasks benefit from human expertise when interacting with AI support.
She emphasizes that relying on AI when interviewing skills are not yet solid is a risky proposition. Instead, the focus should be on developing these foundational skills, with AI serving as a powerful augmentative tool for those who already possess a strong command of continuous discovery principles.
Torres concludes by advocating for a vision of “expert + AI” synthesis that balances speed and quality, cautioning against the long-term costs of outsourcing core customer understanding processes.
📝 About This Content
This article is based on insights shared by Teresa Torres on LinkedIn.
📅 Originally posted on December 2, 2025 | View original post on LinkedIn →