A story crossing a language. A person reclaiming their voice. Technology that helps us be heard.
See it taking shape ↓
Customers can take the next step during the conversation; teams spend more time on the cases that need their judgment.
Vimeo embedded ElevenLabs Dubbing in its video platform. Customers can generate a new language track, edit the text and approve the result without leaving their existing workflow.
Company-published case study · 7 Sep 2026 · customer-reported usage, not an audited company revenue figure
Revolut used ElevenLabs voice technology and deployed engineering while retaining its orchestration and business logic.
An internal prototype required production-scale testing, monitoring, turn taking and payment-data controls.
January 2026 customer case. Comparison includes queues and does not adjust for issue complexity.
Explore two different journeys: a conversation that completes a task, and a piece of content that finds a new audience.
The agent needs to hear the request, handle pauses and respond at the right moment. Good voice AI starts with a conversation that flows.
Which appointment? Which Friday? Connect the request to the information the business needs before taking action.
Connected tools can check availability and update a booking. The workflow needs clear permissions and a reliable response when a tool fails.
Confirm the time and next step. When the request needs a person, transfer the context so the caller can continue the conversation.
Review conversations, test difficult requests and track whether the customer’s task was completed. A pleasant voice is one part of a working service.
A person calling about a card payment needs an answer about that account. Someone booking care needs an appointment in the provider’s system. ElevenLabs started with generated speech, then built tools that connect speech with business actions. Its agent platform lets a team configure knowledge, connect channels, define guardrails, test behavior and monitor conversations. Signet reads the shift as a change in the unit of value: the customer buys a working interaction, with the voice as one part of that interaction.
The company describes three connected layers: research, product and deployment. Researchers develop audio models; product teams build the applications around them; engineers work with organizations to adapt models and agents to specific conditions. That last layer includes language, industry, brand and interaction norms. It explains why a Korea account executive needs to understand the product and the customer’s work, rather than deliver the same demonstration to every prospect. This is Signet’s reading of the model, grounded in the deployment description and current posting.
Revolut first built internally. Its case describes the work needed to combine speech recognition, a language model and speech generation with real-time turn taking, testing, monitoring and payment-data controls. It then evaluated external platforms. In production, Revolut retained its orchestration and business logic while using ElevenLabs for voice quality, responsiveness and reliability. The useful distinction is ownership: connecting to a platform can still leave the customer in charge of the rules that decide what an agent may do.
The same story reports first-phase coverage for more than four million customers in the UK and Europe. Agent tickets took less than five minutes; previous agents’ average ticket time was at least eight times as long. That comparison includes queueing and does not adjust for issue complexity. It therefore describes the rollout’s reported operating experience, rather than a universal productivity multiplier. An applicant can explain why the customer chose the platform while preserving those boundaries around the result.
MasterClass gives a different picture of deployment. Its On Call product required an 18-month collaboration between ElevenLabs’ deployed engineering team and the customer’s product and engineering teams. The work covered dataset curation, model training and quality assurance. More than 75% of On Call users interacted through voice, according to the January 2026 case. This is a usage share for that product. The lesson Signet takes is that voice quality and the customer’s intended experience are developed together through concrete preparation and review.
In the CEO’s April 2026 conversation, small teams and working beside customers are organizational choices, not just sales language. The careers page emphasizes asynchronous documentation, individual ownership and a role-dependent interview process. For a candidate, Signet’s preparation suggestion is to describe a complete piece of work: the initial request, the people involved, the changes made and how you knew it was ready. That connects the company’s stated culture with the practical task of putting an agent or a piece of content into use.
ElevenAgents is the business-interaction platform. ElevenCreative is the production workspace for creators, marketers and media teams. ElevenAPI exposes audio models to developers building their own services. These are separate entry points around related capabilities: generated speech, transcription, music, dubbing and agents. The distinction helps an applicant ask which workflow a customer is buying. A contact-center project, a localized video and a developer integration involve different review steps, even when they use the same underlying speech technology.
The agent workflow extends from a response to a controlled action. A team gives the agent knowledge and procedures, connects channels and business tools, runs simulations and automated tests, then monitors behavior after deployment. The product page connects voice with chat and email rather than treating a call as an isolated interaction. Signet’s reading is that the useful demonstration should include what happens after the answer: a changed record, an appointment, a completed request or an informed transfer to a person.
Klarna’s February 2026 case places ElevenAgents at the first line of US phone support. All 35 million US customers could access the assistant, with resolutions up to ten times faster for queries handled by the agent. Straightforward questions about products, payment status and next steps are the described starting point; access to a human remains part of the design. The planned global expansion to 114 million customers is a separate next step. Eligibility to use the assistant and actual agent usage are different measures.
Vimeo embeds dubbing inside an existing video platform. Customers create a new language track, edit text and approve the result without moving to a separate production workflow. Its September case reports 1.4 million minutes dubbed since January 2026, across more than 25,000 videos. More than 27% of enterprise accounts had tried AI dubbing. These are platform usage measures. They show distribution through a customer’s product; they do not convert directly into ElevenLabs revenue or the audience reached by each localized video.
Dubbing v2 conditions on the original performance so tone and delivery can carry across languages. The developer route puts that capability into a customer’s product, while the creative workspace allows review and editing. Productions adds a network of human specialists for transcripts, subtitles, dubs and audiobooks. Its current freelance opening asks for accurate transcripts and timed subtitles that follow formatting rules. That human work is a distinct part of the product story: language quality must survive the transition from an appealing sample to publishable material.
Eleven v4 and its Turbo variant launched on 28 September 2026. Both support more than 90 languages. The release describes improved expressive delivery, speaker consistency and controllable audio tags, with Turbo optimized together with ElevenAgents. Its approximately 150-millisecond median time to first speech comes from September tests with network latency removed and WebSocket streaming for Turbo. Signet keeps this model-level measurement separate from end-to-end call delay. Integration, tool actions and the caller’s context remain work the deployment has to handle.
May’s company announcement restates year-end 2025 at $350M. April and September values preserve strict lower bounds; September is the company’s Seoul-event statement. These are dated disclosures, not a continuous growth record.
| Date | Value | Basis |
|---|---|---|
| 2025-12 | $350M | Year-end ARR, stated in May 2026 |
| 2026-04 | >$500M | ARR exceeded this threshold by April 2026 |
| 2026-09-03 | >$600M | Company statement at the Seoul event |
ElevenLabs · 디지털데일리 (Digital Daily, 구아현 기자)
Funding rounds and employee tenders are separate transaction types; the latter sells existing shares. These are dated disclosures, not a continuous growth record.
| Date | Value | Basis |
|---|---|---|
| 2025-01 | $3.3B | Series C financing valuation |
| 2025-09 | $6.6B | Employee tender valuation |
| 2026-02 | $11B | Series D financing valuation |
| 2026-09-30 | $22B | Employee tender valuation |
ElevenLabs · ElevenLabs (company blog)
ElevenLabs announced its Korea launch in November 2025 and named collaborations with MBC C&I, ESTsoft, Krafton and SBS. The announcement establishes local relationships, not identical products or outcomes for every partner.
ElevenLabs disclosed a $500M Series D at an $11B valuation. A financing valuation describes that transaction; it does not establish revenue, profitability or a future return.
ElevenLabs announced that ARR crossed $500M by April 2026, after ending 2025 at $350M. These are annualized recurring revenue measures, not audited full-year revenue totals.
ElevenLabs has several commercial routes. A creator buys generation capacity and production tools; a developer buys API access and usage; an enterprise buys platforms, contractual service terms and deployment support. The company’s pricing pages separate self-serve subscriptions from custom enterprise arrangements. Signet reads the common thread as recurring use. A sample earns attention, but a customer continuing to publish content or operate a support workflow is the commercial behavior that turns audio technology into a repeat purchase.
The May 2026 pricing change introduced pay as you go and lowered self-serve API and agent rates. It gives Starter agent usage as $0.08 per minute, compared with the earlier $0.10. Higher-volume customers can negotiate an enterprise plan. That price is a dated plan example, rather than a universal contract rate. Signet’s reading is to compare the purchased minutes with useful work: requests resolved, content approved and integration sustained. The customer’s evaluation includes the work needed to supervise the agent or review its output.
The February 2026 Series D announcement gave year-end 2025 ARR as more than $330M. The May announcement subsequently described that year-end value as $350M and said ARR had exceeded $500M in the first four months of 2026. At the Seoul executive event on 3 September, the company said ARR exceeded $600M. The chart uses the later year-end statement and preserves the lower bounds on April and September. ARR is the annualized scale of recurring business, not revenue recognized in the calendar year.
In September 2026 enterprise represented 55% of revenue. The employee-tender announcement also said agent conversations exceeded fifteen million per week and ElevenAgents ARR had grown more than threefold since February. These measure different things: the mix of total revenue, weekly operating volume and growth within one product. An applicant should preserve those denominators when explaining the company. Signet sees the connection as enterprise demand making deployment and ongoing operation commercially important, rather than treating every user of a generated voice as an enterprise customer.
Series C raised $180M at a $3.3B valuation in January 2025. A September 2025 employee tender used a $6.6B valuation. February 2026 Series D raised $500M at $11B; a September 2026 tender valued the company at $22B and covered $300M of existing employee shares. The chart distinguishes funding rounds from employee liquidity transactions. A valuation establishes the agreed price in that transaction. It is a separate measure from ARR, cash generated by operations or a future listing outcome.
The business also carries execution risks. Signet reads the product and customer evidence as three linked tests: retain a reason to buy despite model competition and lower usage prices; make actions, permissions and human escalation reliable; and adapt content rights, language and data handling to each customer. A Korea enterprise account can involve both commercial judgment and deployment constraints. The current AE posting consequently pairs account growth with deep product understanding, while Productions asks human specialists to keep the published language accurate.
Leads the company’s direction and business. Previously worked at Palantir.
Leads research and engineering. Previously worked on machine learning at Google.
“As AI becomes more intelligent, communication will start to become the limiting factor.”
Mati Staniszewski · Signed investor announcement · 2026-05-05
Mati Staniszewski and Piotr Dąbkowski grew up together in Poland. The official founding account connects their experience of rough dubbing in American movies with a later ambition to carry expression across languages. Mati previously worked at Palantir; Piotr previously worked at Google. They founded ElevenLabs in 2022 and launched its first voice model in January 2023. The origin matters because it begins with the listener’s experience. The research problem is how speech can preserve identity and feeling, not simply pronounce the words on a page.
Mati is co-founder and CEO, responsible for the company’s direction and business. Piotr is co-founder and CTO, responsible for research and engineering. Their February 2026 letter describes an expansion from speech into transcription, music, dubbing and conversational models. The company’s later product description organizes that work into Agents, Creative and API. Signet reads the founders’ choice as building both the audio foundation and the applications that make it useful; it creates a need to coordinate research decisions with customer-facing work.
Ashley Kramer joined as Chief Revenue Officer on 2 September 2026. Her previous role was leading enterprise sales at OpenAI, following senior sales, marketing and product roles at GitLab, Alteryx and Tableau. Her announced remit covers global revenue, the partnership ecosystem and direct work with customers and partners. This places the Korea commercial opening within a company investing in enterprise distribution. It does not establish a reporting line between every regional seller and the CRO; the posting itself names the Revenue, Japan and Korea team.
Ethan Tandowsky joined as CFO on 8 September 2026 after serving as Adyen’s CFO. The announcement assigns him financial infrastructure and investment across research, product and international expansion. At that point ElevenLabs described a global team of 800 and ten million conversational agents launched. Those figures refer to the worldwide team and agents created or launched, respectively. They are useful signs of organizational complexity, but neither is a measure of the Korea team or the number of customers actively operating an agent.
Sangwon Hong is identified in the May 2026 interview as the Korea and Japan GTM director overseeing both markets. He contrasts Japan’s proof-of-concept work in conservative industries with Korea’s faster uptake among startups and digital-native companies. He describes adoption as changing the customer’s operating model and points to manuals and round-the-clock operation as work that must be designed. That is a concrete local responsibility: understand what the customer needs to change before a technically strong voice system becomes something the business can run.
The careers page makes first principles, quality and owning outcomes explicit. It describes documentation for asynchronous collaboration, a hybrid office setup and remote hiring support. Interviews typically include screening, virtual conversations that may include an assessment, a hiring-manager conversation and a founder or Bar Raiser round; the actual sequence varies by role. Signet’s reading for an applicant is practical: prepare evidence of judgment and completion, and ask how the team divides responsibility when a customer’s request crosses sales, engineering and human content review.
Korea’s current openings connect enterprise voice adoption with human review of transcripts and subtitles.
Names shown were identified in public records reviewed on 2026-09-30. Counts cover only the people listed, not total headcount. Grouping does not establish reporting lines.
ElevenLabs announced its Korea launch in November 2025, naming work with MBC C&I, ESTsoft, Krafton and SBS. The company connects Korea with content production and enterprise voice use. In the May 2026 interview, the Korea and Japan GTM director described early uptake among startups and digital-native businesses, then a push toward larger enterprise customers. Signet reads the local opportunity as two kinds of work: help organizations adopt a new customer interaction, and help language-sensitive content pass human review before it is used.
A Korean project reported in April 2026 restored the voice of the late radio DJ Kim Gwang-han for a broadcast with prior family consent. In May, Hong described Spoon Labs reducing production from four to seven months to a matter of hours. The latter is the director’s account of a specific customer’s process, not a guaranteed production time. These examples put rights, delivery and review beside generation quality. They help explain why the local work includes both commercial relationships and localization or transcription functions.
The employer board read on 8 October 2026 contained two Korea-eligible postings among 139 worldwide listings. Account Executive, Korea is full-time and remote within Korea, in Revenue, Japan and Korea. It asks for seven to ten years of SaaS or technology sales, English and Korean fluency, product intuition and proactive account building. Exposure to AI, generative AI, language-model products or API platforms is described as ideal. The duties center on new accounts, use cases and customer value across support, agents and in-app assistants.
Transcription and Subtitling Specialist is a part-time freelance opening in Productions, offered across several locations including Korea and executable remotely worldwide. It asks for native or near-native fluency in a supported language, prior transcription or subtitling experience, familiarity with relevant tools and accurate independent delivery on time. Experience with ElevenLabs or AI-assisted tools, language-service providers, localization or media, and a linguistics or related background are bonuses. The work is claimed and completed through the platform with flexible workload and task-based compensation.
At the September 2026 Seoul event, APAC head Eugene Liang said the company planned more investment in Korean research, product, engineering support and Korean-language technology. Korea data residency was included in the roadmap. The statement concerns expansion plans. The current board’s two openings identify what a candidate can apply for today. Signet keeps those views together without turning a planned engineering expansion into a presently advertised Korea engineering job or treating a regional data-storage roadmap as an already completed local deployment.
The manually observed South Korea LinkedIn view on 8 October showed ten associated profiles. Visible headlines included GTM, sales development, Korea enterprise voice-agent work, dubbing, localization coordination and transcription. Associated profiles may include contractors, and function categories overlap. The visual therefore groups observed work rather than counting a verified employee organization or drawing reporting lines. For an applicant, Signet’s suggestion is to ask how an account, an agent deployment and a reviewed content deliverable move between those functions in practice.
Current Signet catalog · 2026-10-19
Compare the full-time sales job with the task-based freelance engagement.
v4 and Turbo launched on September 28.
The $300M employee tender uses a $22B transaction valuation.
Banner Health’s initial deployment schedules care and transfers to people.
September 2026 brought senior appointments on both the commercial and financial sides. Ashley Kramer became CRO on 2 September and Ethan Tandowsky became CFO on 8 September. The announcements describe an enterprise revenue share of 55%, global revenue and partnership work, and investment decisions across research, product and expansion. For a candidate, Signet reads these appointments as a sign that customer distribution and operating discipline are growing alongside audio research. They describe responsibilities being added to the company, rather than a promise of a particular reporting structure or hiring outcome.
On 28 September the company released Eleven v4 and v4 Turbo. Its announcement ties expressive speech to controllable tags, speaker consistency and agent latency. Two days later it announced a $300M employee tender at a $22B valuation, alongside a description of more than fifteen million agent conversations a week. The product launch and share transaction answer different questions. Signet reads them together as the current chapter: improve how the interaction feels, broaden enterprise operation, and provide liquidity to people who already own shares.
The 2 October Banner Health announcement describes an initial live deployment for primary-care scheduling. Agents schedule, reschedule and cancel appointments in the electronic medical record and transfer to a person with context when human support is needed. Reception, released on 16 September, offers smaller businesses call answering and calendar booking. Vimeo’s September case shows dubbing within a video platform. These examples expand the range of concrete workflows. Signet’s reading is to judge the role by the customer work it helps complete, then distinguish what is live, what is a reported result and what remains a planned expansion.
The board snapshot dated 8 October provides the candidate’s immediate next step: compare the Korea AE role with the freelance Productions engagement, read each full employer description and prepare examples that match the stated requirements. Those openings are separate from the company’s broader international expansion plans and recent executive appointments.
ElevenLabs builds AI audio tools for speech generation, dubbing and conversational agents, helping people and businesses create and use voice across languages.
Korea’s current openings connect enterprise voice adoption with human review of transcripts and subtitles.
Employer postings were checked on 2026-10-08. Availability can change. Mock applications request a review from Signet.
Follow the company from its vision through products, evidence and people to the Korea team and current roles. Use the table of contents to move within the page, and expand the sections for more context.