Discussion about this post

User's avatar
Deborah Carver's avatar

I would be inclined to agree... if the output from generic AI tools in any way matched the level of data quality and research that good human strategists produce.

But when I use the tools, as many others have discovered, it's very clear that LLMs aren't very good at either math or detailed analysis of complex datasets.

Yes, AI could potentially help get to designs faster but the point of strategy is risk mitigation and differentiation, and LLMs are by definition lowest common denominator consensus tools. They are good at high level summary but not much else. The tools can help with the construction and framing of original research, but trusting an LLM with the research itself requires a different level of oversight.

Expand full comment
Karo (Product with Attitude)'s avatar

This resonates a lot: "Living strategies evolve in real time, informed by what prototypes, user feedback, and data are telling us. Strategy is no longer fixed; it’s dynamic."

I think you'll enjoy the post about dynamic user models I published today: https://karozieminski.substack.com/p/user-personas-are-dead-ai-powered

Expand full comment

No posts