In this clip, Devan Batavia shares a quick perspective on founders using AI for strategy. Everyone is using LLMs like Claude and Gemini to build financial models, organize CRMs, and make strategic decisions.
The problem is not that the AI is giving wrong answers. In fact, it is often technically correct. But it completely lacks the broader context of your business.
For example, if you ask an AI how to codify recurring versus one-time revenue in a CRM, it will give a standard, accurate answer. But what it cannot see is your strategic intent. Why are you taking one-off deals in the first place? Is a pilot deal actually a closed deal, or is it still part of your sales pipeline?
The same thing happens with financial modeling. You can ask Claude to build two scenarios: one for going cash-flow positive and one for raising a Series A. The models might look great, but if you do not include variables like revenue concentration risk or potential deal delays, the model is essentially useless.
The old saying "garbage in, garbage out" has not changed just because the technology got smarter. AI cannot see your North Star unless you explicitly show it. If you want accurate strategic output, you have to provide the full picture, not just isolated data points.
What has your experience been with using LLMs for high-level business strategy? Have you run into situations where the AI gave a technically correct answer that was strategically wrong for your project?
View Reddit by JL-Enjoy-The-Work – View Source