Key Takeaways

  • Enterprises spend roughly $6 on outsourced human services for every $1 they spend on software tools.
  • Selling software seats caps your revenue at the tool budget, while selling finished outcomes unlocks the entire services budget.
  • AI companies like Sierra start by assisting human operators, then transition to running entire workflows end to end.
  • Software gross margins of 80% or higher remain intact by shifting from many human operators to automated execution with human judgment only on edge cases.
  • Founders can capture these budgets systematically using The Co-Pilot to Autopilot Outcome Capture Model.

The Co-Pilot to Autopilot Outcome Capture Model

1. Identify the 1:6 Tool-to-Service Ratio

Map a category where customers spend $1 on software tooling but $6 on outsourced human service providers to execute workflows (e.g., $2K on accounting software vs. $15K on accountants).

2. Deploy as a Co-Pilot to Enter the Judgment Loop

Introduce an AI workflow tool alongside existing human operators to handle intelligence tasks while observing and capturing human judgment, taste, and edge cases.

3. Transition to Autopilot Selling Outcomes

Transition from seat-based tool pricing to outcome-based pricing (e.g., price per resolved support ticket, price per closed book) as agent capabilities reach parity.

4. Invert Labor Composition for Software Margins

Shift operations from 'lots of humans, little AI' to 'lots of AI, little humans,' maintaining human escalation loops only for rare judgment calls while retaining 80%+ gross margin software unit economics.

When This Works (and When It Doesn't)

This framework applies directly when building vertical AI applications in large services markets with verifiable, repeatable outputs like customer support, accounting, compliance, and software implementation. The model works because the customer already has a dedicated line item for external labor, making the purchase decision a simple cost comparison against an existing vendor.

It fails in categories where the output is subjective or where liability cannot be cleanly bounded. If a customer cannot define what a good outcome looks like, they will refuse outcome-based pricing and demand human accountability. Traditional agencies also struggle to run this playbook backward. Adding AI tools to a human consultancy rarely changes the firm's cost structure because the incentives remain tied to billable hours rather than software efficiency.

What to Do With This

Look at your product roadmap this week. If you are selling a B2B SaaS tool for $100 per seat per month, find the agency or contractor your customer hires to actually run that tool.

Calculate their annual bill. If your customer pays you $5,000 a year for software but pays an agency $40,000 to manage the work, scrap your next seat-tier pricing update. Package the finished output instead. Offer to deliver the completed work directly for $20,000, keep human operators in the loop to review the edge cases, and let AI handle the underlying execution.