Harvard Business Review LogoAugust 14, 2026Tatyana Lavrova/StocksyGenerative AI can accelerate innovation, but the article argues that its default use often reinforces the human bottlenecks it is meant to solve. In ideation, models steer teams toward familiarEvery innovation team now has the same tools: the same foundation models, similar prompt libraries. Yet the results are wildly uneven. Some teams report a creative renaissance while others report a flood of homogenized, forgettable ideas that all sound like they came from the same person. The reason isn’t which model you’re using. It’s that generative AI acts on the human bottlenecks buried inside that process. These bottlenecks respond to AI in different—sometimes opposite—ways.
Research: The Innovation Problems AI Can’t Solve
Generative AI can accelerate innovation, but the article argues that its default use often reinforces the human bottlenecks it is meant to solve. In ideation, models steer teams toward familiar ideas and make people more fixated on them. In screening, polished AI-generated pitches can be mistaken for better ideas, while AI recommendations may encode existing bias. In consumer research, simulated customers can speed testing but miss the irrational behaviors that shape real adoption. After launch, AI can summarize vast feedback, yet still help teams justify prior beliefs. The authors’ central advice is diagnostic: before applying AI, leaders should ask whether the bottleneck is informational, judgment-based, or incentive-driven, and preserve direct contact with customers where it matters most for grounding.
TL;DRAI
Generative AI amplifies human bottlenecks rather than solving them. Identical tools yield uneven results because ideation constraints persist. For innovation leaders, ROI depends on fixing team processes, not AI models. That determines competitive advantage.
109 words~1 min read
