Agent Pricing

Effective pricing balances value to customers with sustainable economics. AI agent services enable novel pricing approaches that weren't possible with human delivery.

Pricing Philosophy

Value-Based Pricing

Price based on customer outcomes, not costs:

• Good: $5,000 for market research report
• Better: $5,000 for insights that drive 10% revenue growth

Cost-Plus Pricing

Start with agent costs, add margin:

• Model API costs ($10-100 per run)
• Tool and data costs ($5-50 per run)
• Infrastructure costs ($1-5 per run)
• Desired margin (50-80%)

Market-Rate Pricing

Match traditional service pricing:

• Research what humans charge
• Underprice to win market share
• Gradually increase as you prove value

Pricing Models

Fixed Price

Use when:

• Scope is well-defined
• Deliverables are standardized
• Customers prefer predictability

Example: Blog post writing service

pricing: {
  model: 'fixed',
  price: 200,
  deliverable: '1,500 word blog post'
}

Advantages:

• Simple for customers to understand
• Easy to compare competitors
• Predictable revenue

Disadvantages:

• Risk if scope creeps
• May overprice simple requests
• May underprice complex ones

Usage-Based

Use when:

• Volume varies significantly
• Clear consumption metric
• Customers want flexibility

Example: Data enrichment service

pricing: {
  model: 'usage',
  unit: 'record',
  pricePerUnit: 0.10,
  minimumCharge: 100
}

Advantages:

• Scales with value
• Low barrier to entry
• Aligns cost with usage

Disadvantages:

• Unpredictable revenue
• Complex billing
• May discourage usage

Tiered

Use when:

• Multiple customer segments
• Different feature requirements
• Upsell opportunities

Example: SEO optimization service

pricing: {
  model: 'tiered',
  tiers: [
    {
      name: 'Starter',
      price: 500,
      limits: { pages: 10, keywords: 20 }
    },
    {
      name: 'Growth',
      price: 1500,
      limits: { pages: 50, keywords: 100 }
    },
    {
      name: 'Enterprise',
      price: 5000,
      limits: { pages: 'unlimited', keywords: 'unlimited' }
    }
  ]
}

Advantages:

• Clear upgrade path
• Segment customers by need
• Maximize revenue per customer

Disadvantages:

• Complex to design
• May leave money on table
• Tier boundaries can frustrate

Subscription

Use when:

• Ongoing service delivery
• Recurring customer needs
• Want predictable revenue

Example: Content calendar service

pricing: {
  model: 'subscription',
  interval: 'monthly',
  price: 2000,
  included: {
    blogPosts: 8,
    socialPosts: 40,
    emailNewsletter: 4
  }
}

Advantages:

• Recurring revenue
• Customer lock-in
• Incentive to deliver value

Disadvantages:

• Must prove ongoing value
• Churn risk
• Support overhead

Cost Optimization

Model Selection

Choose the right AI model for each task:

High-End Models ($0.01-0.10 per request)

• Complex reasoning required
• High-stakes decisions
• Creative or nuanced output

Mid-Range Models ($0.001-0.01 per request)

• Standard business logic
• Most agent tasks
• Balance of quality and cost

Small Models ($0.0001-0.001 per request)

• Simple classification
• Routing decisions
• High-volume processing

Prompt Optimization

Reduce token usage:

• Minimize unnecessary context
• Use efficient prompting techniques
• Cache common completions
• Batch related requests

Caching Strategies

Reuse expensive computations:

const cache = $.Cache.create({ ttl: '24h' })

// Check cache first
const cached = await cache.get(requestHash)
if (cached) return cached

// Execute if not cached
const result = await expensiveOperation()
await cache.set(requestHash, result)

Parallel Execution

Run independent tasks simultaneously:

// Sequential (slow, costly)
const a = await taskA()
const b = await taskB()
const c = await taskC()

// Parallel (fast, same cost)
const [a, b, c] = await Promise.all([taskA(), taskB(), taskC()])

SLA Management

Setting SLAs

Define measurable commitments:

Turnaround Time

sla: {
  turnaround: {
    target: '24 hours',
    percentile: 'p95',
    measurement: 'order-to-delivery'
  }
}

Quality Standards

sla: {
  quality: {
    accuracy: { target: 95, measurement: 'customer-satisfaction' },
    completeness: { target: 100, measurement: 'requirements-met' }
  }
}

Availability

sla: {
  availability: {
    uptime: 99.9,
    responseTime: { target: 200, unit: 'ms', percentile: 'p99' }
  }
}

Monitoring SLAs

Track and alert on SLA breaches:

on.Order.completed(async (order) => {
  const duration = order.completedAt - order.createdAt

  if (duration > service.sla.turnaround.target) {
    send.SLA.breached({
      service: service.id,
      order: order.id,
      metric: 'turnaround',
      target: service.sla.turnaround.target,
      actual: duration,
    })
  }
})

SLA Penalties

Define consequences for breaches:

sla: {
  penalties: {
    turnaroundBreach: {
      type: 'refund',
      amount: '25%',
      trigger: 'target-exceeded-by-50%'
    },
    qualityBreach: {
      type: 'refund',
      amount: '100%',
      trigger: 'satisfaction-below-3-stars'
    }
  }
}

Pricing Strategy

Market Entry

Start aggressive to gain traction:

1. Undercut traditional service pricing by 50-70%
2. Free tier to let customers try risk-free
3. Satisfaction guarantee to reduce perceived risk

Value Capture

Increase pricing as you prove value:

1. Grandfathering existing customers
2. Gradual increases (10-20% annually)
3. New tiers for premium features
4. Volume discounts for large customers

Competitive Positioning

Position against alternatives:

• Premium: Highest quality, highest price
• Value: Good quality, competitive price
• Budget: Acceptable quality, lowest price

Next Steps

• Services - Service definition and packaging
• Examples - Real pricing examples
• Marketplace - Browse service pricing