Sierra Pricing Strategy
From Bret Taylor: He saved OpenAI, invented the Like button, and built Google Maps • Expert
You're building an AI agent platform for customer service. Your agents can handle customer calls and chats autonomously, resolving issues without human intervention. Early pilots show that your agents can successfully resolve 50-90% of customer service interactions.
You need to decide on a pricing model. You're seeing three common approaches in the market:
- Seat-based pricing (traditional SaaS): Charge per user/agent seat, regardless of usage
- Token/usage-based pricing (popular in AI): Charge per API call or token consumed
- Outcomes-based pricing: Charge based on measurable business results
You have some important context:
- A typical call center phone call costs between $10-20 (mostly labor)
- When your AI agent successfully handles a call, that's a 'call deflection' - the customer never needed a human agent
- You can measure whether your agent actually resolved the customer's problem
- Your product is autonomous - it does the job, not just helps humans do it faster
However, outcomes-based pricing is harder to implement. It requires sophisticated measurement, contract negotiations are more complex, and it's not the industry norm. You'd be going against the grain.
Your investors and board are pushing for the 'safe' choice - usage-based pricing is trendy in AI and easy to implement.
Sign in to submit your response
Get feedback on your approach and see how others solved it.
Continue with Google"I'd make the product better by adding more features and improving the UI. Maybe do some marketing too."
"The core problem is differentiation, not features. I'd start by identifying what job customers are hiring this for that we could do 10x better..."