How to use AI as a sparring partner in your ideation process

Over the past six months, the author worked with 18 product teams using AI tools and large language models (LLMs) for customer discovery and testing, with three teams notably outperforming the others. The key distinction was not access to superior tools but the effective integration of LLMs to enhance already well-understood decisions rather than relying solely on AI-generated solutions. Successful teams used AI to refine their judgment, while less successful teams generated more output with less impact due to over-reliance on AI without critical human assessment. The top teams adopted a "Solo first, AI second" approach, where human judgment led the ideation process, and AI was used as a complementary tool to challenge assumptions and surface overlooked opportunities. A structured four-phase ideation process, including prerequisites, deep problem understanding, solo ideation followed by AI enhancement, and evaluation, was essential for integrating AI effectively. LLMs, as statistical pattern matchers, supported structured thinking and human judgment but required clear context and guidance to be valuable in product innovation, highlighting the irreplaceable role of product leadership in achieving strong outcomes.