Sierra's High-Touch AI Model

Sierra is winning by acting less like a self serve software vendor and more like a modern BPO with AI at the center. In practice, that means Sierra engineers wire the agent into a customer’s help center, CRM, billing stack, and phone flows, then tune real conversations until the bot can resolve a large share of tickets. That makes early deployments work better, but it also makes growth depend on people as much as product.1 2 3

• This model fits the current state of AI support. The category has grown through white glove deployments where vendors write custom integration code, build workflows, and move from feature request to production in 2 to 4 weeks, because enterprise support environments are messy and rarely plug in cleanly out of the box.1
• Sierra has used that hands on approach to land large enterprise brands like ADP, Weight Watchers, and Redfin, and paired it with fast growth, from about $20M ARR in October 2024 to about $100M in October 2025, then to an estimated $200M by May 2026. The tradeoff is that each new account likely carries meaningful service load.2 4
• Incumbents attack the opposite weak point. Intercom bundles AI into a full customer service suite with ticketing, knowledge base, workflows, reporting, and a 99 cent per resolved ticket Fin layer, while Zendesk is pushing no code AI agents inside its existing platform. Those products are easier to turn on, even if they may be less customized at launch.3 5

The path forward is a race to turn custom work into reusable product. The companies that win this market will keep enough engineers close to customers to make automation actually work, then steadily package that learning into templates, integrations, and testing tools so each new deployment needs less human labor than the last.1 3 5