Compounding Engineering Through Reusable Prompts
by Dan Shipper on July 17, 2025
Dan Shipper's company Every operates at the bleeding edge of AI adoption, showing how organizations can transform their operations to maximize productivity and creativity with AI tools. Their approach offers a blueprint for future-focused companies.
Core Organizational Principles
-
Dedicated AI Operations Role
- Employ a head of AI operations who focuses exclusively on building prompts and workflows
- This person identifies repetitive tasks across the organization and automates them
- Separating automation work from day-to-day operations ensures it actually happens
- The role requires someone who:
- Has process orientation skills
- Understands the craft being automated
- Enjoys tinkering and building with AI
-
Compounding Engineering
- For every unit of work, make the next unit easier
- Example: Instead of writing PRDs manually each time, create a prompt that converts rough thoughts into polished PRDs
- Build libraries of prompts that continually refine workflows
- Store these in GitHub repositories where teams can share and improve them
-
Multi-Agent Approach
- Use different AI agents for different tasks based on their strengths
- Engineers maintain relationships with 3-5 different agents simultaneously
- Different agents have different "personalities" and strengths for different tasks
- This creates a "team of Avengers" effect where each agent contributes unique capabilities
-
Code-Free Development
- Engineers don't manually write code; they manage agents who write code
- The workflow shifts to:
- Writing requirements (often with AI assistance)
- Reviewing code generated by AI
- Providing feedback and refinement
- This requires strong programming knowledge but eliminates manual coding
-
Behavioral Change Management
- Implement systems that encourage AI tool adoption
- Example: "Did you put this through the prompt yet?" becomes a standard question
- Create feedback loops where AI usage is visible and celebrated
- Focus first on teams already excited about AI adoption
Organizational Indicators of AI Success
-
CEO Adoption is the #1 Predictor
- If the CEO actively uses AI tools daily, company-wide adoption follows
- CEOs who use AI can set realistic expectations and drive excitement
- Without CEO adoption, initiatives typically fail or face unrealistic expectations
-
Effective Companies Create Visibility
- Hold weekly meetings where people share prompts and use cases
- Send weekly emails showing AI usage stats across the organization
- Highlight early adopters who discover new workflows
- Create forums where AI innovations are rewarded and shared
-
Focus on Augmentation, Not Replacement
- Successful companies focus on doing more with existing teams
- The goal is enabling teams to "go further faster" with the same budget
- This creates positive incentives for adoption rather than fear
-
Generalist Skills Become More Valuable
- AI enables people to work across multiple domains effectively
- Management skills become universally important as everyone becomes a "model manager"
- The ability to evaluate output quality, provide vision, and know when to dive into details becomes critical