AI start-ups: Don’t sell picks and shovels, dig with them

🌈 Abstract

The article discusses the changing landscape of startups and entrepreneurship in the era of AI. It argues that the traditional "picks and shovels" approach of building tools to support other businesses may not be the best strategy in the current AI-driven environment. Instead, the article suggests that founders should consider building full-stack, tech-enabled service businesses that leverage AI at their core, rather than just selling AI-powered software tools.

🙋 Q&A

[01] The Changing Startup Landscape

1. What are the key challenges facing founders in the current AI landscape?

• The rapid development of new AI foundation models and demos is exciting, but the velocity of new technology, lack of clarity around legal issues, general accessibility of the models, and unique commercialization challenges make it difficult for founders to know how to build startups in this new environment.

2. Why does the author think the traditional "picks and shovels" approach may not work this time?

• The author argues that the playbooks that have served startups well over the last 15 years in areas like B2B software development may not be the ones that will help them win in the AI era. The landscape is fundamentally changing, and founders should consider alternative strategies.

[02] Rethinking the Startup Approach

1. What alternative startup models does the author suggest founders consider?

• The author suggests founders should consider building full-stack businesses or a new generation of tech-enabled services, rather than just selling AI-powered software tools. Examples include:
  • A firm built from the ground up with AI tools for tasks like financial statement analysis, auditing, and reporting
  • A new law firm, real estate brokerage, consulting firm, or healthcare practice that leverages AI at its core

2. Why does the author believe this full-stack, service-oriented approach may be more viable in the AI era?

• The author argues that with AI, the need for many historical processes and roles may be eliminated, creating an opportunity for startups to build businesses with AI at the core and compete with legacy organizations that may be slower to fully adopt and understand AI's potential.

3. What challenges have historically made full-stack, service-oriented startups unpopular with founders and VCs?

• Founders have believed that VCs don't like to invest in tech-enabled services businesses, and full-stack startups are harder to execute, have lower margins, and require more people instead of pure software.
• However, the author suggests these objections may no longer hold if AI can dramatically improve the margins and workforce requirements of these types of businesses.

4. What is the author's key advice for founders in the AI era?

• Don't wait for an ecosystem to form around your startup - start building and executing your own full-stack, AI-powered solution. Being just another tool provider may not be a great moat in this new world.