7 Golden Rules for Generative AI Apps: A Playbook from Early Winners - Menlo Ventures

🌈 Abstract

The article discusses the emergence of the first generation of AI-native enterprise applications, which have disrupted traditional enterprise software by leveraging generative AI to unlock new capabilities and markets. It outlines the key strategies and case studies that have enabled these trailblazing companies to achieve rapid growth and success.

🙋 Q&A

[01] Menlo's Seven Golden Rules

1. What are the seven key strategies outlined in the article that have enabled the success of the first wave of AI-native enterprise apps? The seven key strategies are:

• Displace services with software
• Target work that is high-value, high-volume, or facing labor shortages
• Seek pattern-based workflows with high engagement and usage
• Unlock proprietary data
• Embrace zero marginal cost creation
• Build where incumbents aren't, can't, or won't
• Win with compound AI systems rather than models

2. How do these strategies enable AI-native apps to succeed?

• Displacing services with software allows them to target larger markets previously dominated by professional services.
• Focusing on high-value, high-volume, or labor-shortage areas maximizes the impact of AI.
• Targeting pattern-based workflows with high engagement creates a usage moat.
• Unlocking proprietary data builds a data moat.
• Zero marginal cost creation unlocks new content and asset generation opportunities.
• Building in areas where incumbents are weak or slow-moving allows them to establish a foothold.
• Compound AI systems leverage multiple components to create more durable solutions.

[02] Case Studies

1. What are the key insights from the case studies presented?

• Co:Helm used AI to automate prior authorization, a manual professional service, and expand into a much larger market.
• Abridge and Observe.AI targeted high-value and high-volume work respectively, leveraging AI to boost productivity.
• Eve built a usage moat by becoming a critical daily workflow tool for lawyers, rather than relying on data or network effects.
• Eleos unlocked proprietary data by transcribing and structuring patient-therapist conversations.
• Typeface embraced zero marginal cost content creation to unlock new enterprise use cases.
• Sana, SmarterDx, and HeyGen identified areas where incumbents were weak, incapable, or underestimated the opportunity.
• EvenUp won by building a compound AI system rather than a single model.

2. How do these case studies illustrate the application of Menlo's seven golden rules? The case studies provide real-world examples of how the seven strategies have enabled these AI-native apps to succeed, such as:

• Displacing services with software (Co:Helm)
• Targeting high-value and high-volume work (Abridge and Observe.AI)
• Building usage moats (Eve)
• Unlocking proprietary data (Eleos)
• Embracing zero marginal cost creation (Typeface)
• Building where incumbents are weak, incapable, or underestimate the opportunity (Sana, SmarterDx, HeyGen)
• Winning with compound AI systems (EvenUp)