In an official blog post published on 30 October 2026, Sierra said its annual recurring revenue passed $400 million this week, adding more than $150 million in the third quarter alone, more than it added in all of last year. Sierra has been live for a little over two and a half years and now serves nearly half of the Fortune 50. These figures are self-disclosed by the company, with no profit, customer count, or retention rate published alongside them, so the pace of growth and its composition need to be read together.
Read the pace first, then the composition
The timeline Sierra gave is this: it took six months to go from $100 million to $200 million, and only five months to go from $200 million to $400 million, so growth is accelerating. The blog attributed growth to three changes: customers are willing to let agents handle more, and more heavily regulated, work; agents have moved from one-off support conversations to interactions lasting weeks or even months that directly produce revenue; and Ghostwriter, the tool for building and improving agents, has lowered the bar for going live. The blog gave one example: a large US retailer with more than $40 billion in annual revenue went from its first meeting to a live agent in 27 days. That example shows how fast a leading customer can move, not how fast every customer can, and the blog gave no median or overall distribution.
Outcome-based pricing bets on long-running interaction
The key to the Sierra model is not the revenue figure itself but how it charges. The blog restated that customers pay for outcomes, not for tokens consumed, and listed outcomes including a loan being refinanced, an insurance claim being paid, a sale closing, and a customer being retained. Outcome-based pricing only works if agents really can finish work across systems and conversations, and if the context left by each interaction can be reused by the next one. Sierra also said it is developing an open standard called Poppy with more than 50 enterprises, including Meta, Instinct, Rocket, Shopify, Stripe, and Walmart, so companies can control in fine detail how personal agents interact with their brands. For enterprise buyers evaluating this kind of product, the real questions are not whether an agent can answer questions, but which outcomes count as billable, how outcomes are audited, and who bears the cost when something goes wrong. Passing $400 million in ARR shows that enough large companies are willing to pay for this model, but it has not yet answered questions about the profitability and long-term retention of the agent business.
