Ask These Questions Before Starting An AI Startup

Fischer says founders should consider how AI could reshape strategy, products, and teams, including over a two- or three-year horizon. He cautions that enterprise buyers may adopt AI quickly too, using agents to speed purchasing or building software in-house, so startups should not assume incumbents will lag.

He asks whether software will become so easy to generate that businesses build it on demand, or whether AI will help teams create exceptional products that raise the quality bar. He also weighs AI-native startups against established companies adapting their products and teams, and says the right answer may vary by industry and change as models improve.

More capable agents will need secure access to data and clear boundaries between personal and workplace information. Fischer asks how users can know an agent serves their interests, and suggests independent AI audits as one possible way to verify company commitments. For defensibility, he points to proprietary industry knowledge and hard problems such as infrastructure, energy, manufacturing, and chips.

Fischer urges founders to consider what society needs alongside what customers will buy, especially as AI may transform work and concentrate wealth. In the Q&A, he says defensibility matters most for companies meant to last, while founders seeking short-term growth may have different priorities. He also recommends curating a diverse information diet to explore ideas before committing to a company.