Selling AI Systems vs Automations (How we Scaled to 110k+/Mo)

AI Automation Agency vs. Productized AI Systems: A Detailed Comparison

Key Concepts:

• AI Automation Agency: Provides custom workflow automations for clients' existing processes.
• Productized AI Systems: Sells pre-built AI systems designed to deliver specific business outcomes.
• AI SaaS: Selling access to an AI system as a service, typically with a recurring subscription.
• Human-in-the-Loop: Incorporating human decision-making within an automated workflow.
• MVP (Minimum Viable Product): A basic version of a product used for testing and gathering feedback.

Business Model Comparison: Metrics and Definitions

The video compares two AI business models: an AI automation agency and selling productized AI systems.

• AI Automation Agency:
  • Margins: ~40%
  • Net Revenue (17 months): Mid-6 figures
  • Team Size: 16 employees/contractors
  • Focus: Custom workflow automations for existing client processes.
• Productized AI Systems:
  • Margins: ~85%
  • Net Revenue (8 months): Approaching AI automation agency revenue, with exponential growth in the last 3-4 months.
  • Team Size: 4 employees/contractors
  • Focus: Pre-built AI systems delivering specific business outcomes.

The speaker emphasizes that while the AI automation agency had higher initial margins, scaling required hiring more people, reducing margins. The productized AI systems model demonstrates higher scalability and profitability due to its pre-built nature.

Upsides of Each Business Model

• AI Automation Agency:
  • Fast time to revenue: Can start immediately and get paid quickly.
  • Rapid learning: Direct client interaction provides valuable experience.
• Productized AI Systems:
  • Scalability: Build once, sell multiple times.
  • Higher margins: Reduced reliance on custom builds and extensive client interaction.

Challenges of Each Business Model

• AI Automation Agency:
  • Custom builds: Each client requires unique, from-scratch automations.
  • Difficult scoping and pricing: Understanding client processes is time-consuming.
  • High-touch service: Requires significant client communication and involvement.
  • Messy business processes: Often requires establishing or improving processes before automation.
  • Maintenance: Systems can break and require ongoing maintenance.
  • Scaling through hiring: Requires strong management and client relationship skills.
• Productized AI Systems:
  • Time-consuming development: Building valuable, consistent systems takes time and iteration.
  • Domain expertise: Requires deep understanding of a specific niche.
  • Slower time to revenue: Building a product takes time before sales can begin.

Key Skills for Each Business Model

• AI Automation Agency:
  • Hiring and managing people.
  • Client relationship management.
• Productized AI Systems:
  • Product development.
  • Distribution and go-to-market strategy (marketing).
  • Domain expertise in a specific industry.

Automation Types Suitable for Each Model

• AI Automation Agency: Any type of process automation (internal operations, sales, marketing, HR, etc.).
• Productized AI Systems: Primarily sales, marketing, and HR/recruiting (internal operations are often too unique).

AI Automations vs. AI Systems: Definitions and Differences

The speaker defines the key differences between AI automations and AI systems:

• AI Automations:
  • Customized for each business's unique processes.
  • Automate components of larger workflows (tasks).
  • Deliver efficiency wins.
  • Medium business impact.
  • Based on existing customer processes.
  • Can often run autonomously.
• AI Systems:
  • Pre-built systems that work for different companies with the same setup.
  • Automate end-to-end workflows (jobs).
  • Deliver specific business outcomes or contribute to KPIs.
  • High business impact.
  • Based on pre-established processes.
  • Generally require human-in-the-loop decision-making.

Example:

• AI Automation: Keyword research automation (improves efficiency within SEO).
• AI System: SEO e-commerce AI system (automates the entire SEO workflow, including research, content generation, publishing, and analytics).

The speaker's company uses AirTable interfaces to build these systems, allowing human users to execute different tasks and iterate on outputs.

Components of AI Systems

1. User Interface: (e.g., AirTable, Lovable, Bolt, custom code) for human interaction.
2. Database: (e.g., AirTable, Superbase) to store data from automation executions.
3. Multiple Automations: Strung together to form the system (backend).

Ingredients for AI Systems

• Experience and data from a specific industry.
• Pre-built solution (some customization may be involved).

Selling AI Systems: Three Business Models

1. Traditional Agency Route:
   • Present as a traditional agency, using the AI system in the backend to deliver outcomes.
   • Price like a traditional agency (high retainer, commission).
   • Advantage: Sustainable with few leads, can improve the system while making money.
   • Scalability: Limited due to high-touch service.
2. Out of the Box Model:
   • Sell the AI system template to the client (like old software days).
   • Charge a high implementation fee and potentially a low recurring fee.
   • Advantage: More scalable, clients feel they own the solution.
   • Disadvantage: Requires setup and training, gives away the template.
3. AI SaaS:
   • Sell access to the system (subscription-based).
   • Requires low recurring fees and a large customer base.
   • Advantage: No ongoing work, most scalable model.
   • Disadvantage: Risky, requires capital and strong go-to-market experience.

The speaker recommends starting with the agency or out-of-the-box model before transitioning to SaaS, as it allows for iteration and validation.

Building an AI System: A Roadmap

1. Pick an Industry Niche: Play to your strengths (experience, passion).
2. Pick a Business Outcome: Focus on lead generation or recruiting (high-value outcomes).
3. Choose a Path:
   • Traditional Agency Approach: Systemize and automate existing service delivery.
   • AI Automation Agency First: Identify repeatable automations from custom projects.
   • Product First Approach: Build an MVP straight away (requires product and go-to-market experience).

Conclusion

The video provides a detailed comparison of two AI business models, highlighting their respective advantages and challenges. The speaker advocates for a strategic approach to building and selling AI systems, emphasizing the importance of niche focus, business outcomes, and a phased transition towards a SaaS model. The key takeaway is that while AI automation agencies offer a quick start, productized AI systems provide greater scalability and profitability with careful planning and execution.