Cresta
Last reviewed
Jun 4, 2026
Sources
27 citations
Review status
Source-backed
Revision
v1 ยท 2,445 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 4, 2026
Sources
27 citations
Review status
Source-backed
Revision
v1 ยท 2,445 words
Add missing citations, update stale details, or suggest a clearer explanation.
Cresta (Cresta Intelligence, Inc., cresta.com) is an American company that builds generative AI software for contact centers. Its platform combines real-time agent assist (live, on-screen suggestions delivered to human agents during phone calls and chats), conversation intelligence and analytics, automated quality management and coaching, and AI virtual agents that handle customer conversations without a human present. Cresta was founded in 2017 and spun out of the Stanford Artificial Intelligence Laboratory; its co-founders are Zayd Enam, Tim Shi, and the computer scientist Sebastian Thrun, who served as chairman. The company is headquartered in the San Francisco Bay Area and had raised more than $270 million across four venture rounds by late 2024, reaching a valuation of $1.6 billion.[1][2][3]
Cresta's core idea is "human-centric AI": using machine learning to study what the best-performing agents say and do, then surfacing that expertise to every agent in real time, rather than replacing human staff outright. Over time the product line broadened from agent assistance into a unified platform spanning analytics, live guidance, and full automation.[3][4][5]
Cresta grew out of doctoral research at Stanford. Zayd Enam, who held a bachelor's degree in electrical engineering and computer science from the University of California, Berkeley, began a PhD at Stanford in 2015 under Sebastian Thrun, the director of the Stanford AI Lab who had previously led Google's self-driving car project, co-founded Google X, and co-founded the online-education company Udacity. At Stanford, Enam met fellow PhD student Tim Shi, an early researcher at OpenAI. The two left their doctoral programs to commercialize their work, and Thrun, who had advised Enam, joined as co-founder and chairman of the new company.[1][3][6]
The company was incorporated in 2017. Enam became chief executive officer and Tim Shi became chief technology officer. Cresta's early thesis was that contact-center agents vary enormously in effectiveness, and that AI could close that gap by detecting "the psychology, behavior, and language that top contact center agents use" and then coaching every agent toward those patterns in real time.[1][6][3]
| Co-founder | Role | Background |
|---|---|---|
| Zayd Enam | Co-founder, CEO (2017-2023) | PhD candidate at Stanford AI Lab; BS in EECS, UC Berkeley |
| Tim Shi | Co-founder, CTO | PhD candidate at Stanford; early OpenAI researcher |
| Sebastian Thrun | Co-founder, Chairman | Stanford AI Lab director; founder of Google X, Udacity; led Google's self-driving car project |
Cresta operated in stealth for roughly its first three years. It emerged publicly on February 3, 2020, announcing $21 million in total funding to date. That figure included a $15 million Series A round; the financing was led by Andreessen Horowitz and Greylock Partners, with participation from angel investors including Sun Microsystems co-founder Andy Bechtolsheim, Mark Leslie, and Vivi Nevo. At launch, Cresta said Fortune 500 companies were already using the product, and that in lab testing the technology had roughly doubled the throughput of partner sales teams.[1][6][7]
In March 2021, Cresta closed a $50 million Series B round led by Sequoia Capital, with participation from Greylock Partners, Andreessen Horowitz, Allen & Company, and Porsche Ventures. Sequoia's Carl Eschenbach joined the board, which by then also included former AT&T executive John Donovan alongside chairman Sebastian Thrun. The company said its revenue had quadrupled over the prior year.[8][9]
On March 17, 2022, Cresta announced an $80 million Series C round led by Tiger Global at a $1.6 billion post-money valuation, roughly quadrupling its value in a year. The round was notable for drawing strategic investment from contact-center and communications platform vendors: Genesys, Five9, and Zoom all participated, as did J.P. Morgan and CarMax, alongside returning investors Sequoia Capital, Greylock Partners, Andreessen Horowitz, and Porsche. Cresta said it had tripled revenue over the prior year and reported a net revenue retention rate of 210%.[2][10][11]
On November 19, 2024, Cresta announced a $125 million Series D round, bringing total funding to more than $270 million. The round was led by new investors WiL (World Innovation Lab) and QIA, with participation from Accenture, EnvisionX, LG Technology Ventures, Qualcomm, and Workday Ventures, plus returning backers Andreessen Horowitz, Greylock Partners, J.P. Morgan, Sequoia Capital, and Tiger Global. The company said it had nearly quadrupled its annual recurring revenue and nearly doubled its customer base over the prior two years. Cresta's Series D press release did not disclose a new valuation; the most recently announced figure remained the $1.6 billion set at the Series C.[12][13][3]
| Round | Date announced | Amount | Lead investor(s) | Notable participants | Valuation |
|---|---|---|---|---|---|
| Series A (stealth launch, $21M total) | Feb 3, 2020 | $15M Series A | Andreessen Horowitz, Greylock | Andy Bechtolsheim, Mark Leslie, Vivi Nevo | Not disclosed |
| Series B | Mar 2021 | $50M | Sequoia Capital | Greylock, a16z, Allen & Co., Porsche Ventures | Not disclosed |
| Series C | Mar 17, 2022 | $80M | Tiger Global | Genesys, Five9, Zoom, J.P. Morgan, CarMax, Sequoia, Greylock, a16z, Porsche | $1.6B post-money |
| Series D | Nov 19, 2024 | $125M | WiL, QIA | Accenture, Qualcomm, Workday Ventures, LG Technology Ventures, EnvisionX, a16z, Greylock, J.P. Morgan, Sequoia, Tiger Global | Not disclosed |
In May 2023, Cresta named Ping Wu as chief executive officer. Wu, a former Google executive, had co-founded Google's Contact Center AI solution in 2017 and overseen conversational-AI products including Cloud Dialogflow, Cloud Translation, and Speech, and had also helped build Google Cloud's AI Platform products such as Vertex AI and AutoML. Wu had worked with Cresta since 2021 and had been serving as interim CEO before the permanent appointment was announced on May 19, 2023. Co-founder Zayd Enam stepped down from the CEO role, which the company said followed a record fourth quarter, and moved into an advisory capacity. Sebastian Thrun remained chairman and Tim Shi remained as a co-founder in the technology organization.[14][15]
Cresta's offering is built around a single platform that the company frames as three pillars sharing common data, models, and governance: conversation intelligence (analysis), agent assist (augmentation of human agents), and AI agents (automation). The product set has evolved over time, and historical product names such as Cresta Director (manager coaching) and Cresta Insights (analytics) have been folded into the broader platform.[4][5][16]
Agent Assist is Cresta's original product. It listens to a live call or reads a live chat and gives the human agent real-time guidance on screen, including suggested responses, objection-handling prompts, knowledge answers, hints, checklists, compliance reminders, and after-call summaries. The aim is to shorten ramp time for new agents and lift conversion, customer satisfaction, and first-contact resolution by surfacing proven behaviors in the moment. Adjacent capabilities include automatic call summarization and note-taking, AI-assisted coaching, and "smart compose" suggestions for written replies.[4][16][5]
Conversation Intelligence analyzes 100% of customer conversations (rather than a sampled fraction) to surface trends, score quality, automate compliance checks, and capture the "voice of the customer" at scale. It underpins automated quality management (QM) and coaching workflows, letting supervisors identify what drives good and bad outcomes across large agent populations.[4][5][17]
Cresta's AI Agent is a virtual agent that can handle customer conversations autonomously across voice and chat, escalating to a human when needed. The company positions its AI Agent as part of the same platform as its human-agent tools, so that conversation data, knowledge, and analytics are shared between automated and human handling, and so that a conversation handed off from a bot to a person carries context with it.[4][5][16]
Cresta Opera is a self-service "command center" that lets contact-center managers configure the AI without writing code. Through Opera, non-technical users can build rules that recognize conversational behaviors and cues and trigger actions across the platform, driving personalized coaching, automated QA and compliance, and real-time assist. It is the interface through which customers tailor Cresta's models to their own business.[18][16]
Cresta's systems are built on large language models and other natural language processing techniques, fine-tuned on contact-center conversation data. The company combines speech recognition (to transcribe live calls), retrieval-augmented generation (to ground answers in a customer's knowledge base), and models tuned for the domain to produce low-latency suggestions during a live interaction.[19][20][5]
On June 21, 2023, Cresta announced Ocean-1, which it described as the first foundation model purpose-built for the contact center. Rather than training a model from scratch, Cresta built Ocean-1 by combining strong base models, citing OpenAI's GPT-4, Anthropic's Claude, and the open-source Falcon-40B, with contact-center domain data and alignment techniques such as self-instruct and reinforcement learning from human feedback (RLHF). Cresta said Ocean-1 was designed for three goals: better out-of-the-box performance on contact-center tasks, better instruction-following learned from feedback, and lower latency and cost for serving real-time applications at scale. The company later reported that a small, roughly 7-billion-parameter Ocean model could outperform GPT-4 on the retrieval-augmented question-answering that powers its Knowledge Assist feature while being far more cost-effective to run.[19][21][20]
Cresta's broader engineering position is that no single model is optimal for every task in a contact center, so it uses a mix of fine-tuned smaller models and larger general models depending on the job, balancing quality, latency, and cost. The company has said it processes very large volumes of conversations through these systems on behalf of enterprise customers.[20][22]
Cresta sells to large enterprises, typically deploying across contact centers with tens to hundreds of agents per customer and pricing through customized annual contracts. By 2024 the company reported that it had crossed $100 million in annual recurring revenue.[12][23]
Publicly named customers across Cresta's announcements and case studies include Intuit, CarMax, Cox Communications, Hilton, Brinks Home, Porsche, Holiday Inn Club Vacations, United Airlines, Alaska Airlines, and Verizon, among others. In one cited case, Brinks Home reported a 50% reduction in quality-management costs, a 75% first-call-resolution rate, and a 30-point increase in net promoter score after deploying Cresta. Cresta has said customers on average see meaningful improvements in customer satisfaction, faster agent onboarding, and reduced time spent on repetitive tasks, though such figures come largely from the company's own reporting and should be read in that light.[2][12][16]
Cresta has expanded internationally, operating offices in the United States (Palo Alto, San Francisco, and New York) and abroad (Berlin and Toronto), and announcing engineering hubs in Romania and India alongside its 2024 Series D. In its 2025 industry recognition, the company was named a Leader in The Forrester Wave for Conversation Intelligence Solutions for Contact Centers.[12][24]
Cresta competes in a crowded market for contact-center and customer-experience software, which observers generally split into a few groups. Large, full-stack contact-center-as-a-service (CCaaS) and workforce platforms, including NICE, Verint, Genesys, Five9, and Talkdesk, offer broad suites (routing, scheduling, workforce management) with AI added on top. A second group consists of AI-native layers that sit on top of an existing CCaaS to provide analytics, real-time coaching, and automation, where Cresta competes most directly with companies such as Observe.AI, Balto, and Cogito. Notably, Genesys, Five9, and Zoom are simultaneously competitors and strategic investors in Cresta following the 2022 Series C.[10][25][26]
A newer axis of competition comes from AI-agent-first companies, most prominently Sierra, the customer-experience startup co-founded by former Salesforce co-CEO Bret Taylor. Industry comparisons frame the two as approaching the same destination from opposite ends: Cresta began with human-agent assistance and conversation analytics and then added autonomous AI agents, whereas Sierra began with autonomous AI agents and later added tools for human agents. Cresta emphasizes that, unlike pure automation vendors, it also offers quality management and coaching for human staff and a shared data foundation linking analytics, live guidance, and automation.[25][27]
Conversational AI, AI agents, natural language processing, large language model, retrieval-augmented generation, speech recognition, Sebastian Thrun, OpenAI