What Is Sierra AI? Bret Taylor's Agent Platform Explained
Sierra is an enterprise platform for building AI agents that talk to your customers. Companies use it to deploy branded agents that resolve support issues end to end across chat, email, voice, SMS, WhatsApp and ChatGPT, rather than routing every conversation to a human queue.
It is one of the fastest-growing enterprise software companies in recent memory, and understanding what it actually sells is useful whether you are evaluating it, competing with it, or just trying to read the AI agent market accurately.
Who builds Sierra
Sierra was co-founded in 2023 by Bret Taylor and Clay Bavor. Taylor is the chairman of OpenAI’s board and was previously co-CEO of Salesforce, CTO of Facebook, and one of the creators of Google Maps. Bavor spent years running Google Labs and Google’s VR efforts. That pedigree matters commercially, because it is a large part of why Sierra landed Fortune 500 buyers before it had a long product track record.
The funding trajectory has been steep. Sierra raised $950 million in May 2026 led by Tiger Global and GV, at a valuation reported around $15.8 billion, up from roughly $10 billion the previous fall. CNBC reported total capital raised approaching $1.6 billion. The company has said it passed $150 million in annual recurring revenue within eight quarters and counts more than 40% of the Fortune 50 as customers.
What the platform actually does
Sierra’s core idea is that a company should have one agent that represents it everywhere, rather than a chatbot on the website and a separate phone tree and a separate email macro system. You define the agent once, and it deploys across every channel at the same time.
Four pieces make up the product:
Ghostwriter generates production-ready agents from material you already have. You point it at standard operating procedures, past conversation transcripts, photos of internal documentation, or a plain-English description of what the agent should do, and it produces a working multilingual agent with guardrails attached. This is the part that compresses deployment time, and it is the reason Sierra can talk about weeks rather than quarters.
Horizon handles what Sierra calls long-horizon planning, where an agent works toward an outcome that unfolds over days or weeks rather than resolving inside a single conversation. Think of a return that requires a shipment, an inspection and a refund, all tracked as one thread.
Insights analyzes the conversation corpus and surfaces what is going wrong, including problems nobody filed a ticket about.
Observability exposes the agent’s reasoning, which tools it called, and how it performed, which is the part enterprise risk teams actually ask about before signing.
Named customers on Sierra’s own site include Rocket Mortgage, Gap, SoftBank, Singtel, SiriusXM, Wayfair, Vanguard, Uber, Santander, GoFundMe and DIRECTV. The concentration in financial services, telecom and retail is not accidental: those are industries with enormous inbound contact volume and expensive human agents, which is where the math works.
How Sierra prices
Sierra charges on outcomes rather than seats or conversations. You pay when the agent resolves something, not when it merely responds. That model has become the category standard for autonomous customer service, and it exists because it makes the buying decision legible to a support VP who already knows their cost per human-resolved ticket.
Sierra does not publish rate cards. Pricing runs through sales, and the number depends on volume, channel mix and how much of your resolution rate the agent is expected to carry. If you are trying to budget, the honest guidance is that this is enterprise contracting, not a plan you sign up for with a credit card.
Who Sierra is genuinely for
Sierra is built for organizations with a real support operation: a contact center, a queue, a cost-per-contact number someone is accountable for, and enough volume that a percentage point of deflection is worth six or seven figures. If you have those things, it is one of the two or three platforms that will end up on your shortlist.
It is oversized for everyone else. If you are a solo operator, a small agency or a team of five, you do not have the contact volume to justify outcome pricing, you do not have the SOPs Ghostwriter needs to ingest, and the procurement cycle alone will cost more than the problem. That is not a criticism of Sierra. It is a different product category.
The tell is what you are trying to automate. Sierra automates conversations your customers start with your company. If what you actually want to automate is your own working day, the inbox triage and the scheduling and the CRM updates and the follow-ups that nobody outside your company ever sees, that is a personal and small-team automation problem, and the tooling looks nothing like a contact center platform. Carly sits in that second category: an AI assistant with its own email address that fires on real events across 260+ integrations, with free Zapier-style workflows and AI agents from $35/month. Worth knowing which of the two problems you have before you take a demo.
Sierra compared to the rest of the category
The closest direct comparison is Decagon, which sells a very similar promise to a very similar buyer and has raised aggressively on the same thesis. Sierra is generally the larger and more enterprise-weighted of the two, with more Fortune 50 logos and a heavier emphasis on the agent-building tooling; Decagon competes hard on speed of resolution and has moved faster in some mid-market segments. We break the two apart in more detail in Decagon vs Sierra.
The incumbent comparison is different in kind. Zendesk and Intercom both now ship AI agents of their own, and for many companies the real decision is not “which AI agent vendor” but “do we bolt an agent onto the helpdesk we already run, or do we put a dedicated agent platform in front of it.” Vendors like Sierra argue the helpdesk-native agents inherit the ticketing model’s assumptions. The incumbents argue they already hold your data and your workflows. Both arguments have merit, and the answer usually turns on how much of your resolution you expect to automate.
FAQ
Is Sierra AI the same as Salesforce?
No. Bret Taylor was co-CEO of Salesforce before founding Sierra, but Sierra is an independent company with its own platform and funding. It integrates with CRMs including Salesforce rather than being part of one.
How much does Sierra AI cost?
Sierra uses outcome-based pricing, meaning you pay per resolved interaction rather than per seat or per message. Rates are not published and are negotiated through sales based on volume and channel mix.
What channels does Sierra support?
Chat, email, voice, SMS, WhatsApp and ChatGPT, with agents deploying across all supported channels from a single definition.
Is Sierra suitable for small businesses?
Generally no. The pricing model, the deployment process and the sales cycle all assume meaningful contact-center volume. Small teams typically get more from general-purpose automation tooling than from a dedicated customer-agent platform.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
